# nimo full documentation > Complete machine-readable copy of the public nimo docs. Canonical docs index: https://heynimo.com/docs Short agent discovery file: https://heynimo.com/llms.txt Agent guide: https://heynimo.com/AGENTS.md Agent skill: https://heynimo.com/skill.md ## Product summary - nimo is an AI website performance agent. - nimo follows one workflow: Find the gap. Explain why. Fix or hand off. Watch the result. - nimo checks Chrome field data when available, monitors uptime, triages regressions and incidents, and helps users decide the first safe action to review or hand off. - nimo connects to Telegram, Slack, MCP clients, Google Search Console, Google Analytics, Ahrefs, Cloudflare, and custom MCP servers when configured. - Cloudflare changes require a signed-in user, explicit approval, and safe actions enabled for the account. - Agentic readiness is optional and advisory. Treat WebMCP and Agentic Browsing findings as readiness signals, not guarantees. ## How agents should use these docs - Start with /docs for onboarding and /docs/mcp for agent setup. - Use the public docs MCP at https://api.heynimo.com/docs-mcp when a client can connect MCP and only needs public documentation. - Use the authenticated nimo MCP only when the user has a nimo account, creates an API key, and wants an agent to read account sites, audits, history, or uptime. - Use each page's Markdown URL when a concise page-level source is enough. - Use this full bundle when a client needs all public docs in one request. - Do not assume access to private account data. Authenticated MCP access requires a user-created API key. - Prefer exact product limits and caveats from the Billing and API keys section. ## Documentation pages ## Quick start nimo follows one workflow: Find the gap. Explain why. Fix or hand off. Watch the result. Start here if you want the shortest path from account to useful answer. Canonical URL: https://heynimo.com/docs Markdown URL: https://heynimo.com/docs.md ### Summary The fastest path for a new user: Find the gap. Explain why. Fix or hand off. Watch the result. Create an account, add a site, run the first audit, connect Telegram or Slack, and ask the nimo agent a focused performance question. ### Key points - Find the gap with CrUX field data, Lighthouse diagnostics, history, and competitor context when useful. - Explain why the gap matters with source labels instead of raw audit jargon. - Fix or hand off one safe first action with owner and acceptance criteria. - Watch the result with reruns, scheduled audits, competitor watches, and uptime checks. - Create an account from the public sign-in flow. - Add one production site URL. - Add important pages from the Pages tab when Page Control is available. - Run the first audit. - Connect Telegram, Slack, or both. - Ask: What should I fix first on mobile, and why does it matter? ### Start in three minutes - Create your account so nimo can save sites, history, alerts, and integrations. - Paste the public URL you want nimo to monitor. The dashboard starts the first audit and turns on uptime monitoring for that site. If the Pages tab is available, add the most important URLs after the first site audit. - Connect Telegram, Slack, or both so nimo can send audit updates and answer follow-up questions in chat. - When the first audit finishes, ask one focused question: What should I fix first on mobile, and why does it matter? ### What happens next - After the first audit, nimo shows a plain-language report with real visitor data from Chrome when available, simulated Lighthouse test data, prioritized recommendations, chat follow-ups, uptime status, audit history, and incident evidence when enough regression evidence exists. - Use the same workflow after every important change: Find the gap. Explain why. Fix or hand off. Watch the result. - Smaller or newer sites may not have enough Chrome field data yet. nimo says that clearly and uses the simulated test as a diagnostic baseline. ### First prompts to try - Ask short prompts and make one decision at a time. - Good prompts: What is the single biggest reason this page feels slow? Which slow page is most likely hurting search traffic or conversions? Compare my last two audits. Did the fix actually help? ### Keep it running - Use Telegram or Slack for chat, audits, site switching, and alerts. - Use MCP when Claude Code, Cursor, Windsurf, or another agent needs to read audit history and run checks. - Connect Google data for search and traffic context. - Connect Cloudflare when you want nimo to suggest edge fixes for images, compression, caching, and headers. ## Set up a site Add a site and run the first audit. nimo will monitor uptime, store history, and use that context when you ask questions. Canonical URL: https://heynimo.com/docs/setup Markdown URL: https://heynimo.com/docs/setup.md ### Summary Explains site onboarding, Page Control for important URLs, URL validation, automatic uptime checks, scheduled audits, and the difference between real visitor data and simulated test data. ### Key points - Use the public production URL visitors use. - Use Page Control for important URLs when the Pages tab is available. - Paid plans can use weekly, daily, or manual audit schedules. - Real visitor data is primary when Chrome has enough traffic for the page. - Simulated test data is diagnostic and useful when field data is missing. ### Add your site - Sign in and choose Add site. Use the public URL visitors use, including the protocol, for example https://example.com. - Start the first audit. nimo checks the URL, pulls available Chrome field data, runs a simulated test, and builds the report. - Read the plain-language verdict first. It answers what is slow, why it matters, and what to do next. - Add production pages when you want search and real visitor context. Staging pages can still be useful for quick simulated checks. ### Choose the right schedule - Paid plans can use weekly, daily, or manual audit schedules. - Use weekly audits when a site changes occasionally. - Use daily audits when the page is tied to paid traffic, high-value SEO traffic, checkout, lead capture, or client reporting. - Run a manual audit any time you make a change and want a before-after check. ### Add important pages - If the Pages tab is available for your site, use Page Control to tell nimo which URLs matter most. - Add pages such as the homepage, pricing page, signup or checkout path, product pages, comparison pages, high-traffic content, and docs. - When discovery is enabled, the Pages tab can suggest same-site URLs from sitemap and crawl results. - Discovered rows stay pending until you confirm and save the page profile. - Each page can have a type, importance, funnel role, owner, device priority, monitoring schedule, and short notes. - Use these fields to route alerts and recommendations to the right person without making every URL look equally important. - Start with public production URLs that affect revenue, signups, search traffic, or support. Private, local, and unsafe targets are rejected by nimo's fetch guardrails. ### Understand the data sources - Real visitor data is Chrome field data. It shows what real visitors experienced and reflects many visits, not one test run. - Simulated test data is Lighthouse-style lab data. It helps diagnose why a page is slow, but it is not the same as what real visitors saw. - Page inspection lets nimo inspect structure, scripts, images, and common performance patterns when answering chat questions. - Google, Ahrefs, Cloudflare, and MCP connections add traffic, SEO, infrastructure, and workspace context. ### Manage site settings - Use site settings to change schedules, run manual audits, compare history, manage important pages when Page Control is enabled, connect integrations, and remove sites you no longer want nimo to monitor. - Keep the first setup simple: add one important page first, then add more money pages or client sites after the recommendations make sense. ## Telegram Telegram is the main place to talk to nimo. Connect it once, then ask questions, run audits, switch sites, and get alerts without opening the dashboard. Canonical URL: https://heynimo.com/docs/telegram Markdown URL: https://heynimo.com/docs/telegram.md ### Summary Short setup and usage guide for Telegram, including account linking, site switching, audits, status checks, new chats, unlinking, and useful questions to ask. ### Key points - Create a Telegram link token from Settings. - Press Start in Telegram to finish linking. - Use /site, /addsite, /audit, /status, /newchat, and /unlink. - Ask short questions and let nimo answer from the active site's report history. ### Connect Telegram - Open Settings in the web app and find Telegram. - Click Connect. nimo creates a private Telegram deep link for your account. - Telegram opens the nimo chat. Press Start to finish linking. - If you have more than one site, use /site to choose which one nimo should answer about. ### Use the commands - /site switches the active site. - /addsite adds a new site from Telegram. - /audit runs an audit for the active site. - /status checks current site and connection status. - /newchat starts a fresh chat thread. - /unlink disconnects Telegram from your nimo account. ### Ask short questions - Telegram works best when the question is direct. - Good prompts: Why did my mobile LCP get worse this week? Run a new audit and tell me if the hero image fix helped. What is the next safest thing to fix? ### When nimo messages you - nimo reaches out for performance regressions, uptime incidents, completed audits, or fixes that need approval. - Uptime notifications are batched and quiet during late hours unless an incident has lasted long enough to be urgent. ## MCP for agents Use the public docs MCP when an agent only needs nimo documentation. Use the authenticated nimo MCP server when an agent needs your performance data. Canonical URL: https://heynimo.com/docs/mcp Markdown URL: https://heynimo.com/docs/mcp.md ### Summary Shows how to use the public docs MCP server, create a nimo API key, configure remote MCP clients with bearer auth, and use the nimo MCP tools for sites, audit reports, history, comparisons, and uptime. ### Key points - Use the public docs MCP at https://api.heynimo.com/docs-mcp when an agent only needs public nimo documentation. - Create an API key in Settings. Keys start with nimo_ and are shown once. - Connect an HTTP MCP client to https://api.heynimo.com/mcp-server. - Send Authorization: Bearer nimo_YOUR_KEY. - Tools include list_sites, get_site, get_audit_report, get_audit_history, compare_audits, run_audit, get_audit_status, and get_uptime_status. ### What the MCP server gives agents - The MCP server lets an agent inspect sites, read audit reports, compare history, start audits, and check uptime. - Ask your coding agent to read the latest nimo report before it changes code. It can target the page, metric, and recommendation that matter most. ### Public docs MCP - Use https://api.heynimo.com/docs-mcp when an agent only needs public nimo documentation. - This server is read-only and does not require an API key. - Tools include list_docs, search_docs, read_doc, and read_llms_full. - Use the authenticated nimo MCP server only when an agent needs account data such as sites, audits, history, or uptime status. ### Create an API key - Open Settings and find API keys. - Create a named key for a client such as Claude Code or Cursor. - Copy it once. nimo shows the key only once, and it starts with nimo_. - Paste it into the MCP client as a bearer token in the Authorization header. ### Connect your client - Claude Code command: claude mcp add nimo https://api.heynimo.com/mcp-server -t http -s project -H "Authorization:Bearer nimo_YOUR_KEY" - Generic MCP config: use url https://api.heynimo.com/mcp-server and header Authorization: Bearer nimo_YOUR_KEY. - For Windsurf, add the same server URL and bearer header in ~/.codeium/windsurf/mcp_config.json. ### Security notes - API keys are shown once. Store them in your agent secret store or local MCP config. - Revoke keys you no longer use. - Each user can have up to five keys. - MCP access requires an active paid or comped plan. - Treat MCP clients like any automation that can read account data and trigger audits. ## Audits An audit is the find-the-gap step. nimo checks how your site performs, explains what matters, and saves the result so future changes can be compared. Canonical URL: https://heynimo.com/docs/audits Markdown URL: https://heynimo.com/docs/audits.md ### Summary Covers full and quick audits as the workflow's find-the-gap step: CrUX field data, Lighthouse lab diagnostics, audit progress, history, and before-after comparisons. ### Key points - Audits combine Chrome field data, simulated lab diagnostics, page inspection, and AI analysis. - Run audits from the dashboard, Telegram, or MCP. - Compare audits before and after fixes to check for improvement. - Use audits to support the workflow: find the gap, explain why, fix or hand off, and watch the result. ### Run an audit - Pick the site from the dashboard, /site in Telegram, or list_sites through MCP. - Start the audit from the UI, /audit, or the run_audit MCP tool. - A full audit pulls available Chrome field data, runs a simulated test, analyzes the result, and creates a report. - A quick audit is useful after a fix when you want a fast comparison. ### Read the results - nimo leads with a plain-language verdict, then shows metrics and recommendations. - LCP describes how long visitors wait before the main content appears. - CLS describes how much the page shifts while visitors try to read or click. - INP describes how responsive the page feels when visitors interact. - Diagnostics cover images, scripts, stylesheets, fonts, third parties, and page structure. - Lighthouse scores can move between runs. nimo focuses on metric values, history, and real visitor impact. ### Compare changes - Run an audit before and after a fix. - nimo can compare two audits and show what improved, what got worse, and what stayed the same. - Good prompt: Compare my last two audits. Which metrics changed enough to trust? ### Use audit history - Audit history helps nimo tell the difference between a one-off slow run and a real trend. - For important pages, run a manual audit before deploying a fix, then run another audit after the change ships. ## Monitoring and alerts Monitoring is the watch-the-result step. nimo keeps watching after the first audit by checking speed on a schedule, watching competitor gaps you choose, checking uptime every few minutes, and telling you when something needs attention. Canonical URL: https://heynimo.com/docs/monitoring Markdown URL: https://heynimo.com/docs/monitoring.md ### Summary Explains the workflow's watch-the-result step: scheduled performance audits, competitor watches, degradation alerts, incident evidence, Linear issues, verification, uptime checks every five minutes, batched notifications, and quiet hours. ### Key points - Scheduled audits build history and detect regressions. - Competitor watches check chosen public URL pairs weekly. - Incident evidence explains meaningful regressions with scoped evidence, owner, confidence, and Linear issue context when enabled. - Uptime checks run every five minutes. - Notifications are batched and quiet during 10pm to 8am Pacific unless urgent. - Use monitoring to watch whether the first safe fix held up over time. ### Scheduled audits - Scheduled audits build the history nimo needs to spot regressions. - Use weekly audits for sites that change occasionally. - Use daily audits for important pages, frequent deployments, paid traffic, checkout, lead capture, or client reporting. - Use manual audits when you make a change and want a before-after check. ### Competitor watches - Competitor watches turn a public comparison into a weekly check. - Run a competitor comparison, choose Watch this gap, then sign in or create an account. - nimo saves the two public URLs and checks the pair weekly. - Weekly summaries tell you whether the competitor pulled ahead, you pulled ahead, the gap is tied, or the check did not have enough data. - When public CrUX history is available, weekly summaries can add one short mobile LCP trend line so you can see whether either side is improving. - Competitor watches store normalized public URLs, domains, status, next run time, last run time, a short latest gap summary, and an optional public CrUX trend summary. - They do not store cookies, request headers, credentials, private pages, or raw Lighthouse blobs. ### Manage competitor watches - Open Watches in the app sidebar to review active and paused competitor watches. - Use Pause when you do not want weekly checks for a pair. - Use Resume to start weekly checks again. - Use Edit to rerun the comparison with the same URLs and decide what to watch next. - Use Delete to stop the watch and remove it from your active list. - Paused and deleted watches do not run new weekly checks. - Delete a watch when the competitor is no longer relevant or when you do not want nimo to keep checking that public URL pair. ### Degradation alerts - nimo compares recent audits with previous results and alerts when metrics move beyond normal noise. - The alert explains what changed and what to check first. - Good prompt: Why did this alert fire, and what changed on the page? ### Incident evidence and Linear issues - When performance incident tracking is enabled, a meaningful regression can become an incident with stored proof. - nimo uses baseline and follow-up evidence to show the affected metric, before-after values, scope, evidence refs, likely owner, confidence, and recommendation context. - Incident evidence can appear on the dashboard, site report, and audit history when enough evidence exists. - It is meant for triage, not for storing raw traces or private request details. - If Linear is configured, you can preview the issue nimo would create for an active incident. - Creating the issue requires an explicit create action, and verification waits for a fresh audit after the task, deploy, or manual check. - Weekly summaries can include unresolved incidents, recoveries, open Linear issues, and changed important-page profiles when reporting is enabled. - Slack and Telegram delivery follows your audit notification preferences. ### Uptime checks - Uptime starts automatically when you add a site. nimo checks every five minutes. - Up means the site responded. - Degraded means the site responded, but slowly. - Down means the site timed out, had DNS trouble, or returned a server error. - A 4xx response counts as reachable because many CDNs and WAFs reject automated requests even when the site is alive. ### Quiet hours - Uptime notifications are batched for 90 seconds so one user does not get a message for every site at once. - During 10pm to 8am Pacific time, non-urgent notifications wait unless an incident has lasted 30 minutes or more. - nimo uses batching and a circuit breaker to avoid false notification storms. ## Reports and sharing Reports are built for the explain-why and fix-or-hand-off steps. nimo tells you what is slow, why it matters, and what to review first. Canonical URL: https://heynimo.com/docs/reports Markdown URL: https://heynimo.com/docs/reports.md ### Summary Describes the explain-why and fix-or-hand-off parts of the workflow: nimo AI reports, prioritized recommendations, platform-aware quick wins, incident evidence, Agentic readiness, competitor proof reports, comparison methodology, six-month CrUX trend history, visual proof screenshots, copyable fix prompts, Markdown and PDF exports, field and lab source labels, public share links, and client-ready summaries. ### Key points - Reports include a verdict, metrics, deep dives, and prioritized recommendations. - Platform-aware quick wins translate a finding into a simple path for WordPress, Shopify, Webflow, Next.js, Cloudflare, or an unknown stack. - Incident evidence explains stored regressions with scoped evidence, metric deltas, likely owner, confidence, and Linear issue context when enabled. - Agentic readiness is optional and advisory; treat WebMCP, llms.txt, and AI-agent interaction findings as readiness work you can opt into. - Competitor proof reports show who wins on LCP, why they win, and the first safe fix to review or hand off. - Visual proof can add stored screenshots when safe, but LCP, source labels, and proof facts still decide the gap. - Suggested and default competitors are optional. The user confirms the final URLs before any comparison runs. - Copy fix prompts turn a safe first fix into a coding-agent handoff. - Markdown and PDF exports include the same public evidence with source labels. - Share links create public expiring URLs for reports and comparisons. - Use report comparisons for client before-after evidence. ### What a report includes - Verdict: a short explanation of the current state of the page. - Metrics: Core Web Vitals and supporting metrics from field and lab data. - Deep dives: images, scripts, fonts, third parties, page entities, search context, and traffic context when available. - Next steps: recommendations ordered by likely user impact, practicality, and safety. - Incident evidence: stored regression evidence when nimo has enough baseline and follow-up data. - Readiness findings: advisory AI-agent readiness checks when the readiness surface is enabled. ### Prioritized recommendations - nimo does not simply repeat every Lighthouse warning. - It ranks work by likely user impact, stack relevance, metric trend, and safety. - A recommendation can be important even when the Lighthouse score looks fine. nimo checks field data first when it exists. ### Platform-aware quick wins - When nimo has a confident stack signal, the report can show one platform-aware quick win for WordPress, Shopify, Webflow, Next.js, or Cloudflare. - Stack detection is best-effort guidance, not a guarantee. Confirm the platform before changing production. - The card keeps the user journey simple: platform, owner, one action, why it comes first, and the confidence note. - Owner labels are explicit: developer, marketer, CMS owner, or hosting/CDN. nimo does not remove plugins, disable apps, edit code, or change Cloudflare settings from public chat. Cloudflare changes require a signed-in user, explicit approval, and safe actions enabled for the account. - Examples: WordPress advice may start with the hero image, plugin scripts, or page cache. Shopify advice may start with image sizing, app embeds, pixels, or theme code. Webflow advice may start with assets, embeds, and first-section interactions. - Next.js advice may start with next/image sizing, font loading, client-only bundle splitting, route caching, or data fetching. Cloudflare advice only points to cache, image delivery, or headers when the evidence supports it and the page is safe to cache. ### Understand field, lab, and missing data - CrUX field data shows real Chrome visits over roughly 28 days. It is not first-party RUM and it is not real-time. - CrUX can lag after a deploy and may use origin-level data when page-level data is missing. - Lighthouse lab data is a point-in-time simulation for the tested URL and device. - Lighthouse is useful for debugging and immediate post-fix checks, but it may not match what real visitors experienced. - If field data fails and lab data passes, fix the field failure first because real visitors are seeing the issue. - If lab data fails and field data passes, treat the lab failure as a diagnostic clue before overriding passing field data. - If CrUX is missing, the page or origin may not have enough Chrome traffic. Use Lighthouse as a baseline, then rerun the same device after each fix. ### Incident evidence - When performance incident tracking is enabled and nimo has enough evidence, a report can show incident evidence for a regression. - The panel focuses on what changed, the metric delta, the affected page or site scope, evidence refs, likely owner, confidence, and recommendation context. - Incident evidence is not a raw trace viewer. It uses stored, scoped evidence so a teammate can understand the regression without exposing cookies, credentials, request headers, raw Lighthouse payloads, or private URLs. - If Linear is configured, nimo can preview an issue from an active incident. - Creating the issue requires an explicit create action, and duplicate issue links are reused when nimo can identify an existing issue. ### Agentic readiness - Agentic readiness is optional and advisory. - It checks whether AI agents can discover and interact with the public site more reliably, using signals such as /llms.txt, WebMCP metadata, declarative forms, AI accessibility, and Lighthouse Agentic Browsing output. - Readiness findings do not mean performance monitoring failed. - Treat them as optional AI-discovery and interaction-quality work, separate from Core Web Vitals, uptime, and incident alerts. - Eligible readiness findings can show a Linear issue preview after opt-in. - nimo keeps the handoff narrow and reuses already ticketed findings instead of creating duplicate work. ### Competitor proof reports - Public competitor comparisons answer the workflow questions before the metric table: find the LCP gap, explain why the faster page wins, and choose the first safe fix to review or hand off. - The leaderboard orders your page and each competitor by LCP, with the source labeled as CrUX field data, Lighthouse lab data, or unavailable. - CrUX field data is the 28-day p75 from real Chrome users. It is the best signal for what visitors experienced, but it can lag after a deploy and may use origin-level data when page-level data is missing. - Lighthouse lab data is a point-in-time test for the exact URL. It is useful right after a fix because you can rerun it immediately. - The proof section compares simple facts such as page weight, JavaScript weight, request count, third-party scripts, TTFB, and render-blocking resources. - The first fix includes an owner, likely impact, and acceptance check so the next step is easy to hand to a developer, marketer, or hosting team. - When safe screenshots are available, visual proof can show your page next to the faster competitor. - When public CrUX history is available, the report shows whether mobile and desktop LCP are improving, flat, or getting slower. - When a comparison mixes CrUX field data and Lighthouse lab data, treat the result as directional. After a deploy, rerun Lighthouse right away and wait 2-4 weeks for CrUX field data to settle. - Public proof reports only include audit facts from public URLs. They do not expose request headers, cookies, secrets, or raw Lighthouse blobs. LCP element snippets are cleaned before display. ### Comparison methodology - Source labels stay visible for every comparison fact: CrUX field data, Lighthouse lab data, URL-level CrUX history, origin-level CrUX history, or unavailable. nimo does not blend field and lab data into one unlabeled score. - Live comparisons run only after the user confirms the URLs and presses Compare speed. Lighthouse lab data is a point-in-time test for the exact URL and device. - CrUX field data reflects the latest available public 28-day Chrome window and can lag after deploys. CrUX trend history uses roughly six months of public history when available. - Competitor watches rerun chosen public URL pairs weekly. - The leaderboard orders pages by measured LCP for that comparison only. It is directional evidence, not a promise of search ranking improvement, conversion lift, Core Web Vitals passing status, or the same order after a later CrUX update or Lighthouse rerun. ### CrUX trend history - CrUX trend history uses public Chrome usage data to show LCP momentum over roughly the last six months. - It is not first-party RUM, it is not real-time analytics, and it may be missing for newer or lower-traffic pages. - The trend section compares the first and latest public LCP values and labels each page as improving, flat, getting slower, or not enough samples. - nimo keeps mobile and desktop trend summaries separate because the same page can move differently on each device. - The report labels whether the trend came from URL-level CrUX history or origin-level CrUX history. - Origin fallback means Chrome did not have enough page-level history. - If a page does not have enough public history, nimo says that plainly and keeps the current Lighthouse and CrUX field labels visible. - Use CrUX trend history to understand momentum, not to prove a fix the same day it ships. For immediate verification, rerun Lighthouse on the same device, then watch CrUX over the next few weeks. ### Visual proof screenshots - Visual proof is a side-by-side screenshot pair from the same kind of Lighthouse lab run used for the comparison. - Screenshots come from a single Lighthouse lab capture per page. Use them as directional evidence, then rerun after a fix. - Visual proof may be unavailable when screenshots are missing, expired, too large, unsupported, or fail safety checks. - Reports use stored PNG or JPEG artifacts from the audit. They do not hotlink competitor screenshots from third-party URLs. - Markdown exports include the visual proof caption only. PDF exports include thumbnails only when the stored image data is safe. - Treat visual proof as the simple screenshot explanation for a measured gap. Use LCP, source labels, proof facts, and a rerun after deploy to decide whether the fix worked. ### Suggested and default competitors - On the public comparison form, nimo can suggest up to three competitor URLs after the user enters their page. - Suggestions are optional. The user can pick a suggestion, edit it, remove it, or ignore suggestions and type competitor URLs manually. - Default competitor presets are editorial starting points used by comparison pages. They prefill known public URLs, but they do not start a live audit by themselves. - nimo only sends the target URL to the suggestion endpoint. Analytics for this flow record counts and source labels, not competitor domains or full URLs. - The public endpoint validates hostnames, rejects private or local targets, rate-limits requests by IP and domain, and returns a manual fallback when suggestions are unavailable. - When Ahrefs-powered organic suggestions are enabled, nimo returns only the normalized public suggestion URLs. It does not expose raw provider payloads or authenticated Ahrefs data. - Before a live comparison runs, the user must confirm one to three competitor URLs by pressing Compare speed. ### Copy fix prompts - When a public comparison has a safe first fix, use Copy fix prompt to copy a concise handoff for Claude Code, Cursor, or another coding agent. - The prompt includes the public URL, winning competitor URL, failing metric, source labels, first fix, sanitized evidence, acceptance criteria, and a rerun instruction. - The prompt tells the agent not to change unrelated behavior and to ask before destructive changes. - The prompt excludes credentials, cookies, request headers, private account data, raw Lighthouse blobs, raw HTML, and URL query strings. - Paste the prompt into your coding tool, review the plan, make the smallest safe change, and rerun the comparison after deploy. ### Export Markdown and PDF - Public competitor comparisons can be copied as Markdown or downloaded as Markdown and PDF. - Markdown is best for tickets, pull request notes, agency task lists, and coding-agent handoffs because it is plain text. - PDF is best for stakeholder or client handoff when the reader needs a compact, readable report. - Both exports include the verdict, compared URLs, leaderboard, proof facts, visual proof caption when available, first fix when available, rerun guidance, generated date, and source labels. - Markdown does not embed screenshots or external image URLs. PDF thumbnails are included only from stored safe screenshot artifacts. - Source labels stay visible in exports because CrUX field data and Lighthouse lab data answer different questions. CrUX shows real Chrome users over a 28-day window; Lighthouse shows an immediate test for the exact URL. - Exports are generated from the public share token on demand. They do not include credentials, cookies, request headers, private account data, raw Lighthouse blobs, raw HTML, or URL query strings. ### Share a report - Use share links to send a report or comparison to a client, teammate, or developer. - Open the report, create the share link, choose the expiry window, and send the link. - The recipient can view the report without signing in and can download the public Markdown or PDF export when exports are enabled. ### Use reports with clients - For client work, save a before audit, make the change, run another audit, and share the comparison. - Good prompt: Summarize this report for a client in five bullets. Focus on business impact and the next fix. ## Chat with nimo Chat is where reports become useful. nimo knows your site context and can answer specific questions without making you translate performance jargon. Canonical URL: https://heynimo.com/docs/chat Markdown URL: https://heynimo.com/docs/chat.md ### Summary Explains how the nimo agent uses audit data, site history, uptime, page inspection, saved persona and focus preferences, web tools, and connected integrations to answer performance questions. ### Key points - nimo chat can use audit data, history, uptime status, page inspection, preferences, and connected integrations. - Settings can persist persona and focus areas so chat answers match the user's role and priorities. - Ask one focused question at a time. - Good prompts ask what changed, what to review first, or whether a fix worked. ### What nimo knows - nimo can use the latest audit and report, audit history, uptime status, page inspection, web tools, saved preferences, and connected Google, Ahrefs, Cloudflare, and MCP integrations. ### Good questions to ask - Ask nimo to make a decision, explain a change, or choose the next fix. - Good prompts: What should I fix first if I only have one hour? Explain the LCP issue like I am sending it to a developer. Did the last deployment make this page faster or slower? ### Set persona and focus - Use the AI tab in Settings to save the role and priorities nimo should assume when it answers. - Persona is a short role label, such as developer or marketer. - Focus areas are the performance topics you care about most, such as Core Web Vitals, checkout conversion, mobile UX, or client reporting. - These preferences shape chat context. They do not replace report evidence, site history, or source labels. ### Run work from chat - Depending on setup, nimo can trigger audits, inspect a page, check uptime, compare audits, and suggest Cloudflare optimizations. - For Cloudflare, chat can inspect settings, suggest changes, and save approval history. Apply and rollback require a signed-in user, explicit approval, safe actions enabled for the account, and the web apply flow; public chat cannot change Cloudflare settings. - When an active performance incident exists, chat can explain the stored incident evidence and the likely next owner. - Linear issue creation and recovery verification require explicit ticket or verification intent, a connected Linear workspace, and the relevant feature gates. - Chat is most useful after an audit because the report gives nimo evidence and avoids generic advice. ### Keep conversations useful - Ask one question at a time. - Name the page or site if you manage several. - Ask for the reason behind each recommendation. - Ask for a smaller next step when an answer feels too broad. ## Integrations Integrations give nimo more context and a place to reach you. Connect only what helps your workflow. Canonical URL: https://heynimo.com/docs/integrations Markdown URL: https://heynimo.com/docs/integrations.md ### Summary Covers Slack, Google Search Console, Google Analytics 4, Ahrefs MCP, Cloudflare MCP, Linear issues, custom MCP connections, and the current status of Notion. ### Key points - Slack adds chat and audit updates in your workspace. - Google Search Console and Analytics add search and traffic context. - Ahrefs adds SEO data through MCP. - Cloudflare adds read-only edge setting review, optimization suggestions, approval history, apply, and rollback through OAuth and nimo web flows. - Linear can receive explicit incident and Agentic readiness issues when enabled and connected. - Custom MCP servers can add narrow workspace tools. - Notion is currently marked coming soon in the public Settings UI. ### Slack - Slack lets you talk to nimo from a DM or channel mention and receive audit updates in chat. - Connect Slack during onboarding or from Settings. - Settings shows connection status, connected workspace, connect, and disconnect states. - Invite nimo to a channel when you want channel replies. ### Google Search Console and Analytics - Google Search Console adds search context. Google Analytics 4 adds traffic context. - Open Settings, choose Google integrations, authorize Google, and pick the account that owns the right properties. - If Analytics data is missing, reconnect Google so nimo can request the Analytics readonly scope. ### Ahrefs - Ahrefs adds SEO data such as domain rating, organic traffic, keywords, backlinks, and top pages. - Create an Ahrefs MCP key in Ahrefs, then paste it into nimo Settings. ### Cloudflare - Cloudflare is used for read-only edge setting review and optimization suggestions by default. - Connect with OAuth. nimo asks for permission to read account, analytics, zone, and zone settings data for diagnostics and suggestions. - Apply and rollback require a signed-in user, explicit approval, and safe actions enabled for the account; public chat and public docs cannot change Cloudflare settings. ### Linear - Linear is used for explicit incident and Agentic readiness issues when the integration is enabled. - nimo can preview a narrow issue from stored evidence, then create the issue only after you explicitly confirm. - If Linear is not connected, nimo can offer a connection link during an explicit ticket flow. - Issue creation requires a connected workspace, a default team, and an explicit create action. ### Custom MCP servers - Use a custom MCP connection when your own tools or workspace data should be available in nimo chat. - Add the server URL, transport, and any required headers in Settings. - Keep custom MCP servers narrow. The best integrations answer one job clearly, such as finding docs, reading a CMS, or looking up workspace release notes. ### Notion - Notion is planned for workspace context, but it is not available in public Settings yet. ## Cloudflare optimization When your site runs behind Cloudflare, nimo can review read-only settings, suggest edge fixes, save your approval, apply approved safe actions when they are enabled for your account, and require separate approval before rollback. Canonical URL: https://heynimo.com/docs/cloudflare-optimization Markdown URL: https://heynimo.com/docs/cloudflare-optimization.md ### Summary Explains Cloudflare OAuth, read-only setting review, the three-audit learning phase, supported edge optimization suggestions, explicit approval, apply, rollback approval, and verification. ### Key points - Connect Cloudflare with read-only OAuth scopes. - nimo waits for a three-audit baseline before suggestions. - Supported fixes include image optimization, Brotli, Early Hints, Auto Minify, and Browser Cache TTL. - Approval saves who approved the change and the test-run details. Apply and rollback are available only through signed-in actions when safe actions are enabled for the account, and Cloudflare must confirm the expected value before success is recorded. ### Connect Cloudflare - Open Settings, go to Integrations, and choose Cloudflare. - Sign in with Cloudflare and allow account read, analytics read, zone read, and zone settings read scopes. - nimo needs enough audit history before it suggests infrastructure changes. ### Learning phase - nimo waits for a three-audit baseline before suggesting Cloudflare changes. - The learning phase helps nimo avoid changing settings because of one noisy run. ### Supported fixes - Image Polish and WebP can improve image delivery at the edge when images are a real bottleneck. - Brotli compression can compress text assets so pages transfer less data. - Early Hints can let browsers begin loading important resources sooner. - Auto Minify can minify HTML, CSS, or JavaScript when it is safe for the site. - Browser Cache TTL can improve repeat visits by caching static assets longer. ### Approve, apply, and roll back - nimo can save that you approved a suggested Cloudflare change so the recommendation history is clear. - Approval prepares an expiring approval request with test-run details; it does not change Cloudflare settings from public chat. - When safe actions are enabled for the account, a signed-in apply action can apply an approved request through backend code. - nimo records apply success only when Cloudflare confirms the desired value. - Rollback is separate: nimo requires a rollback plan, an available rollback reference, and a second explicit approved rollback request before changing settings back. - Rollback success is recorded only when Cloudflare confirms the restored value. - Use `getOptimizationHistory` in chat to see suggested and approved Cloudflare actions. - After an approved safe action changes infrastructure, run or wait for verification to check the result. ### Safety constraints - Cloudflare write actions are unavailable unless safe actions are enabled for the account. - Public docs, public chat, and unauthenticated agents cannot apply or roll back Cloudflare settings. - Every apply or rollback should include scope, approval, rollback or forward plan, Cloudflare confirmation, and verification. ## Billing and API keys Plans control how many sites, audits, chat messages, and AI tokens your account can use. API keys are for agent and MCP access. Canonical URL: https://heynimo.com/docs/billing-and-api-keys Markdown URL: https://heynimo.com/docs/billing-and-api-keys.md ### Summary Summarizes Pro and Business limits, subscription management, usage counters, API key creation, one-time key display, key revocation, and bearer auth for MCP. ### Key points - Pro is $20/month: 10 sites, 50 audits/month, 200 chat messages/month. - Business is $199/month: 50 sites, 200 audits/month, 1,000 chat messages/month. - API keys are for MCP and agent access. They are stored as hashes and can be revoked. ### Plans - Pro costs $20/month and includes 10 sites, 50 audits/month, and 200 chat messages/month. - Business costs $199/month and includes 50 sites, 200 audits/month, and 1,000 chat messages/month. - Business raises the monthly token budget from 500K to 2M and is better for teams, agencies, and accounts that need more sites, audits, or chat volume. ### Usage limits - nimo tracks sites, audits, chat messages, and token usage. - If you hit a limit, upgrade or wait for the monthly reset. - Comped accounts may have limits handled separately. ### API keys - API keys power the nimo MCP server. - Keys start with nimo_, are shown only once, are stored as hashes, can be revoked from Settings, and are limited to five active keys per user. - Bearer auth requires an active paid or comped account. - Create the key in Settings, store it safely, and send it as Authorization: Bearer nimo_YOUR_KEY. ### Keep keys safe - Use one key per agent or environment. - Revoke keys when a laptop, teammate, project, or client no longer needs access. ## Comparison pages # nimo vs PageSpeed Insights A fair comparison of PageSpeed Insights and nimo for testing a page, explaining the result, and choosing the next step. Canonical URL: https://heynimo.com/compare/pagespeed-insights Markdown URL: https://heynimo.com/compare/pagespeed-insights.md Sources checked: May 19, 2026 ## AI summary nimo is useful when performance evidence needs to become a clear next step with source labels and a check-again step. ## Key caveats - Do not claim nimo can promise rankings, conversions, or passing Core Web Vitals. - Do not claim nimo modifies a site by default. Measurement and explanation are default. Cloudflare changes require a signed-in user, explicit approval, and safe actions enabled for the account. - Keep field data, lab data, and unavailable data separate. Do not blend them into one unlabeled score. - Recheck official competitor sources before adding new capability, pricing, packaging, quota, API, or monitoring-limit claims. ## When nimo fits - Use nimo when PageSpeed Insights gives useful evidence, but the next step still needs a plain-language owner, first action, and check-again step. - Use nimo when CrUX field data and Lighthouse lab data disagree and the labels need to stay visible before deciding what to change. - Use nimo when the result needs to become a report for a developer, marketer, CMS owner, hosting team, or agency. ## When PageSpeed Insights fits - Use PageSpeed Insights when the user wants Google's free mobile and desktop diagnostic for a single page. - Use PageSpeed Insights when the user needs a quick Lighthouse report without creating an account. - Use PageSpeed Insights when the user already knows how to turn raw diagnostics into a safe implementation plan. ## Source labels and caveats - CrUX field data is public Google field data from real Chrome usage when enough data is available. - CrUX commonly reflects a previous 28-day collection period, so it is not same-day proof that a deploy fixed a field gap. - Lighthouse lab data is a controlled point-in-time diagnostic run for debugging a specific page. - Lighthouse can disagree with field data because it is a test environment, not the whole audience. - Unavailable means the page or origin did not have enough public field data for the requested view. ## Official sources - [About PageSpeed Insights](https://developers.google.com/speed/docs/insights/v5/about) - [Chrome UX Report data on PageSpeed Insights](https://developer.chrome.com/docs/crux/guides/pagespeed-insights) ## Related nimo pages - [Website speed comparison tool](https://heynimo.com/compare-websites) - [Free Core Web Vitals audit](https://heynimo.com/free-core-web-vitals-audit) - [Reports and sharing docs](https://heynimo.com/docs/reports) # nimo vs GTmetrix A fair comparison of GTmetrix and nimo for deeper reporting, simpler handoffs, and next-step decisions. Canonical URL: https://heynimo.com/compare/gtmetrix Markdown URL: https://heynimo.com/compare/gtmetrix.md Sources checked: May 19, 2026 ## AI summary nimo is useful when performance evidence needs to become a clear next step with source labels and a check-again step. ## Key caveats - Do not claim nimo can promise rankings, conversions, or passing Core Web Vitals. - Do not claim nimo modifies a site by default. Measurement and explanation are default. Cloudflare changes require a signed-in user, explicit approval, and safe actions enabled for the account. - Keep field data, lab data, and unavailable data separate. Do not blend them into one unlabeled score. - Recheck official competitor sources before adding new capability, pricing, packaging, quota, API, or monitoring-limit claims. ## When nimo fits - Use nimo when the user needs a simpler explanation of CrUX, Lighthouse, and mixed-source comparison data. - Use nimo when the useful output is one first action, owner, and check-again step instead of a broad dashboard review. - Use nimo when the result needs to become a shareable report for a developer, marketer, CMS owner, hosting team, or agency. ## When GTmetrix fits - Use GTmetrix when the user wants a familiar performance report with waterfall, video, report history, and filmstrip views. - Use GTmetrix when the user needs scheduled testing, alerts, testing locations, or configurable analysis options. - Use GTmetrix when the team already uses GTmetrix monitoring workflows and wants to keep that dashboard depth. ## Source labels and caveats - GTmetrix describes reports with Lighthouse metrics and audits, Web Vitals, CrUX Real User Metrics, waterfall, video, report history, and filmstrip views. - Use GTmetrix feature claims as sourced competitor strengths, not as proof that every tool runs the same lab conditions. - CrUX field data can lag after a deploy, so it is better for trend context than same-day fix proof. - Lighthouse lab data is a controlled point-in-time diagnostic run and can differ across tools when device, network, location, or analysis settings differ. - This page does not make exact pricing, location, monitoring-limit, API-credit, or packaging claims; recheck official GTmetrix sources before adding them. ## Official sources - [GTmetrix homepage](https://gtmetrix.com/) - [GTmetrix features](https://gtmetrix.com/features.html) - [GTmetrix pricing](https://gtmetrix.com/pricing.html) ## Related nimo pages - [Website speed comparison tool](https://heynimo.com/compare-websites) - [Free Core Web Vitals audit](https://heynimo.com/free-core-web-vitals-audit) - [Reports and sharing docs](https://heynimo.com/docs/reports) ## Organic guides # Why speed tools disagree, and what to trust first Guide for understanding why PageSpeed Insights, Search Console, Lighthouse, and GTmetrix can disagree, plus which result to trust first. Canonical URL: https://heynimo.com/field-vs-lab-core-web-vitals Markdown URL: https://heynimo.com/field-vs-lab-core-web-vitals.md Sources checked: May 20, 2026 ## AI summary Use this guide when someone sees conflicting speed results across PageSpeed Insights, Search Console, Lighthouse, or GTmetrix and needs to know which source to trust first. ## Key caveats - CrUX field data reflects real Chrome usage when enough data is available, but it is delayed and is not first-party RUM. - Lighthouse lab data is a controlled diagnostic for one tested page and device profile; it can disagree with real-user field data. - Search Console groups similar URLs and is meant for site-level troubleshooting, not precise same-day verification of one URL. - GTmetrix is useful lab and monitoring evidence, but its test conditions can differ from PageSpeed Insights, Search Console, and local Lighthouse runs. - Do not claim a fix worked from one source alone. Rerun the lab check after deploy, then watch field data over the next CrUX window. ## Source labels - PageSpeed Insights can show CrUX field data and Lighthouse lab diagnostics when enough public field data exists. - Search Console's Core Web Vitals report comes from CrUX field data and groups URLs by issue and device type. - Lighthouse is a point-in-time lab audit that helps identify likely causes such as render-blocking resources, heavy images, scripts, and layout shifts. - GTmetrix describes PageSpeed Insights as a tool with CrUX real-user data and GTmetrix as lab or synthetic testing in controlled environments. - nimo keeps those labels visible instead of blending field data, lab data, and unavailable data into one unlabeled score. ## How to resolve contradictions - If Search Console fails but Lighthouse passes, treat field data as the user-impact signal and use lab diagnostics to find the likely cause. - If Lighthouse fails but Search Console passes, treat Lighthouse as an immediate debugging clue before overriding passing field data. - If GTmetrix looks good but PageSpeed Insights field data fails, check whether the GTmetrix result is lab evidence while PSI is reporting CrUX field evidence. - If PageSpeed Insights has no field data, use the lab result as a baseline and monitor the page until enough public field data exists. - After a deploy, rerun the same lab check immediately, then wait for field data to update before calling the field gap closed. ## Where nimo fits - nimo is useful when the result needs to become one first action with an owner and check-again step. - nimo is useful when a stakeholder, client, developer, marketer, CMS owner, or hosting team needs a clear explanation instead of raw diagnostics. - nimo is not a replacement for every tool. PageSpeed Insights, Search Console, Lighthouse, and GTmetrix are still useful evidence sources. ## Sources - [Google PageSpeed Insights documentation](https://developers.google.com/speed/docs/insights/v5/about) - [Google Search Console Core Web Vitals report](https://support.google.com/webmasters/answer/9205520) - [Chrome UX Report overview](https://developer.chrome.com/docs/crux) - [Chrome Lighthouse documentation](https://developer.chrome.com/docs/lighthouse) - [GTmetrix FAQ](https://gtmetrix.com/faq.html) ## Related nimo pages - [Run a free Core Web Vitals audit](https://heynimo.com/free-core-web-vitals-audit) - [Compare a page against competitors](https://heynimo.com/compare-websites) - [Read the reports and sharing docs](https://heynimo.com/docs/reports) # Shopify Core Web Vitals first fixes Practical Shopify Core Web Vitals checklist for separating field data from lab diagnostics and reviewing theme, apps, third-party tags, and images first. Canonical URL: https://heynimo.com/shopify-core-web-vitals-first-fixes Markdown URL: https://heynimo.com/shopify-core-web-vitals-first-fixes.md Sources checked: May 20, 2026 ## AI summary Use this guide when someone asks what to fix first for a Shopify Core Web Vitals problem and needs a practical checklist with source caveats. ## Key caveats - Shopify says online stores already include global hosting, CDN, browser caching, gzip compression, image optimization, and file minification behavior. - Shopify says theme code, installed apps, and manually added third-party code are major factors that affect store performance. - Third-party tool recommendations may already be implemented by Shopify or may not apply to Shopify stores. - Do not remove apps, pixels, reviews, chat, or merchandising features without checking business value and ownership. ## Shopify first-fix checklist - Label the failing source first. Owner: SEO or marketer. Check whether the problem is CrUX field data, Lighthouse lab data, a Search Console URL group, or an unavailable field sample before changing the theme. - Review the first visible section. Owner: Theme owner. Check the hero image, video, animation, carousel, and number of homepage or product-page sections before editing lower-page content. - Audit apps and third-party tags. Owner: Store owner. List installed apps, pixels, tag-manager scripts, reviews, chat widgets, and personalization tools. Keep the ones with clear revenue or workflow value. - Fix image sizing before chasing server settings. Owner: Theme owner. Use appropriately sized images and confirm the theme lets Shopify's image CDN serve efficient formats and dimensions for the first contentful view. - Rerun the same check after the change. Owner: SEO or developer. Use the same URL, device, and source label for the after check. Watch CrUX later before declaring the field issue closed. ## Why the platform matters - Shopify already includes hosting, CDN, caching, compression, and image delivery behavior. Many generic recommendations need a Shopify-specific translation before they are worth changing. - Start by labeling field data, lab data, and unavailable data before changing platform settings. - nimo turns that evidence into one owner, one first action, and one check-again step. ## Sources - [Shopify performance overview](https://help.shopify.com/en/manual/online-store/web-performance/overview) - [Shopify improving web performance](https://help.shopify.com/en/manual/online-store/web-performance/improving-web-performance) - [Google PageSpeed Insights documentation](https://developers.google.com/speed/docs/insights/v5/about) ## Related nimo pages - [Run a free Core Web Vitals audit](https://heynimo.com/free-core-web-vitals-audit) - [Understand field and lab disagreements](https://heynimo.com/field-vs-lab-core-web-vitals) - [Compare a Shopify page against competitors](https://heynimo.com/compare-websites) # WordPress Core Web Vitals first fixes Practical WordPress Core Web Vitals checklist for separating field data from lab diagnostics and reviewing cache, theme, plugins, images, and scripts first. Canonical URL: https://heynimo.com/wordpress-core-web-vitals-first-fixes Markdown URL: https://heynimo.com/wordpress-core-web-vitals-first-fixes.md Sources checked: May 20, 2026 ## AI summary Use this guide when someone asks what to fix first for a WordPress Core Web Vitals problem and needs a practical checklist with source caveats. ## Key caveats - The WordPress performance handbook describes caching as the fastest way to improve performance for many WordPress sites, but cache changes still need site-specific validation. - The WordPress Performance Lab plugin is a collection of performance feature projects, not a blanket guarantee that a site will pass Core Web Vitals. - Do not combine multiple cache, minify, image, and script-delay plugins without checking conflicts and rollback paths. - WooCommerce, membership, learning, form, and logged-in flows can need different cache rules than static marketing pages. ## WordPress first-fix checklist - Label field data, lab data, or missing data. Owner: SEO or developer. Do not treat one Lighthouse score as the whole truth. Check whether CrUX field data exists, whether Search Console groups the URL, and what the lab run is diagnosing. - Confirm page caching for public pages. Owner: Developer or host. For mostly static pages, confirm a page cache or host cache is serving HTML without breaking logged-in, cart, form, or personalized flows. - Inspect the LCP element and hero media. Owner: Content or theme owner. Find the main image, heading, video poster, or first content block. Check file size, dimensions, loading priority, and whether lazy loading is delaying above-the-fold content. - Review plugin and builder scripts. Owner: Site owner. List scripts loaded by page builders, forms, analytics, popups, sliders, reviews, ads, and chat. Remove, delay, or scope low-value scripts only after confirming ownership. - Rerun and compare the same source. Owner: SEO or developer. Rerun the same lab check after deploy and compare the metric that failed first. Let CrUX field data catch up before calling a field failure fixed. ## Why the platform matters - WordPress can be fast, but every site has its own host, theme, builder, plugin stack, cache layer, and content model. The first fix should match the bottleneck instead of stacking optimization plugins blindly. - Start by labeling field data, lab data, and unavailable data before changing platform settings. - nimo turns that evidence into one owner, one first action, and one check-again step. ## Sources - [WordPress performance cache handbook](https://developer.wordpress.org/advanced-administration/performance/cache/) - [WordPress Performance Lab handbook](https://make.wordpress.org/performance/handbook/performance-lab/) - [Google PageSpeed Insights documentation](https://developers.google.com/speed/docs/insights/v5/about) ## Related nimo pages - [Run a free Core Web Vitals audit](https://heynimo.com/free-core-web-vitals-audit) - [Understand field and lab disagreements](https://heynimo.com/field-vs-lab-core-web-vitals) - [Read the reports and sharing docs](https://heynimo.com/docs/reports) # Client website speed report template Client-ready website speed report template for agencies that need a short verdict, first fix, owner, share links, Markdown/PDF exports, and a check-again step. Canonical URL: https://heynimo.com/agency-core-web-vitals-report-template Markdown URL: https://heynimo.com/agency-core-web-vitals-report-template.md Sources checked: May 20, 2026 ## AI summary Use this page when an agency, freelancer, or consultant needs a simple client report structure with evidence, owner, first fix, check-again step, and export/share options. ## Key caveats - Client reports should keep CrUX field data, Lighthouse lab data, Search Console groups, and unavailable data visibly separate. - Do not promise ranking, conversion, revenue, or passing Core Web Vitals from one report. - Use lab reruns for immediate before-after checks, then wait for CrUX field data before claiming a field issue recovered. - Public share links and exports should include only report evidence that is safe to show to the recipient. ## Client report structure - Short verdict: one sentence describing the current state and affected page experience. - Source labels: name which facts came from CrUX field data, Lighthouse lab diagnostics, Search Console URL groups, or unavailable data. - Metric snapshot: show LCP, INP, CLS, and TTFB only when the source exists and the metric is relevant. - Proof: show the likely cause, affected element or resource class, and any before-after comparison that exists. - First fix: name the owner, first safe action, and check-again instruction. - Caveats: say what the report cannot prove yet, especially after same-day changes. - Share/export: send a public share link or Markdown/PDF export when the client needs the report outside the app. ## Sample report section - Verdict: Mobile LCP is poor in CrUX field data, and the lab run points to the hero image plus blocking CSS as the first place to inspect. - Source labels: CrUX field data is delayed real Chrome-user evidence; Lighthouse lab data is the immediate diagnostic run for the exact URL. - First fix: Developer or CMS owner should resize and prioritize the hero image, then run the same mobile lab check again. - Acceptance check: Lab LCP should improve on the same URL and device profile; CrUX field data should be reviewed again after the next field-data window. ## Where nimo fits - nimo reports are built for this handoff: verdict, metrics, proof, next action, owner, and verification step. - Public report links let a recipient view a safe report without signing in. - Markdown exports are useful for tickets, pull requests, and coding-agent handoffs. - PDF exports are useful for stakeholders who need a compact client-facing artifact. ## Sources - [Google PageSpeed Insights documentation](https://developers.google.com/speed/docs/insights/v5/about) - [Google Search Console Core Web Vitals report](https://support.google.com/webmasters/answer/9205520) - [Chrome UX Report overview](https://developer.chrome.com/docs/crux) ## Related nimo pages - [Run a free Core Web Vitals audit](https://heynimo.com/free-core-web-vitals-audit) - [Read reports and sharing docs](https://heynimo.com/docs/reports) - [Compare a client page against competitors](https://heynimo.com/compare-websites) - [Understand field and lab disagreements](https://heynimo.com/field-vs-lab-core-web-vitals) ## Current product caveats - Notion is currently marked coming soon in the public Settings UI. - API keys are shown once and require an active paid or comped account for MCP bearer auth. - Cloudflare optimization is advisory by default. nimo can suggest settings and save approval history; apply and rollback require a signed-in user, explicit approval, and safe actions enabled for the account. - Agentic readiness is optional and advisory. Treat WebMCP and Agentic Browsing findings as readiness signals, not guarantees.