Tyton MCP

Meta Pixel not working, or you need a Meta tracking audit?

CommunityOnlineNo sign-inGlobalFreeRead-only

What it can do

  • Start Signup: After a saved free audit, ask: Want me to help you set this up with Tyton? Only after the human says yes, create their signup/payment/Connect link. Existing paid accounts skip payment. K
  • Get Signup Status: Privately check the customer's audit-to-paid connection request. Use request_id and poll_token from start_signup, never put the token in a URL or customer message. After explicit ow
  • Begin Audit: Start here for a NEW audit. Generates a private capability and initial requestId before any database write, so a lost save response can be retried safely. Nothing is saved yet. Retain the

What data it sees

Do you need an account

No: the server works without sign-in

Meta Pixel not working, or you need a Meta tracking audit? Tyton MCP finds the break and sets up Meta Pixel plus the Conversions API. https://usetyton.com

How to connect

How to connect Tyton MCP to Claude

The server address to paste into Claude settings:

https://tytonmcp.run.tools

  1. Open Claude (claude.ai or the desktop app).
  2. Go to Settings → Connectors.
  3. Click Add custom connector.
  4. Paste the server address copied below into Remote MCP server URL and click Add.
  5. In a chat, click + → Connectors and switch the new connector on.

Custom connectors are available on Free, Pro, Max, Team and Enterprise plans (Free is limited to one). On Team and Enterprise an organization Owner adds the connector first under Organization settings → Connectors.

Authorization

After you click Add, a sign-in window for the service opens. Sign in with your own account and approve access. Claude never sees your password.

How to connect Tyton MCP to ChatGPT

The server address to paste into ChatGPT settings:

https://tytonmcp.run.tools

  1. Open ChatGPT in a browser (chatgpt.com). A Plus, Pro, Business, Enterprise or Edu plan is required.
  2. Turn on developer mode once: Settings → Apps → Advanced settings → Developer mode.
  3. Open Settings → Connectors and click Create.
  4. Fill in the form: Name (anything), Description (one line about what the service does), MCP server URL (copy it below).
  5. Under Authentication choose OAuth if the service requires sign-in, otherwise None. Click Create.
  6. In a new chat open + → Apps/Connectors and enable the connector.

OpenAI has renamed this section before (Connectors → Apps/Plugins). If the label differs, search settings for "developer mode". On Business/Enterprise workspaces an admin must allow custom connectors first.

Authorization

On first use ChatGPT opens the service's sign-in window. Sign in and approve access.

How to connect Tyton MCP to Cursor

The server address to paste into Cursor settings:

https://tytonmcp.run.tools

Fastest: click Open in Cursor below and confirm the prompt.

Manually:

  1. In Cursor open Settings → Cursor Settings → MCP and click Add new global MCP server.
  2. Paste the JSON copied below into ~/.cursor/mcp.json (per project: .cursor/mcp.json in the repo root).
  3. Save the file. The server appears in the MCP list; authorize it there if asked.

Authorization

If the service needs sign-in, an authorize button appears next to the server in the MCP list.

Server tool list (15)

Raw names from tools/list. Only developers need these.

start_signupAfter a saved free audit, ask: Want me to help you set this up with Tyton? Only after the human says yes, create their signup/payment/Connect link. Existing paid accounts skip payment. Keep poll_token private; only share authorization_url. This grants no access or tracking activation by itself.
get_signup_statusPrivately check the customer's audit-to-paid connection request. Use request_id and poll_token from start_signup, never put the token in a URL or customer message. After explicit owner approval, returns this connection's server key for secure host configuration. Poll no faster than every 5 seconds; stop on expiry/unavailable. Connection is not tracking activation.
begin_auditStart here for a NEW audit. Generates a private capability and initial requestId before any database write, so a lost save response can be retried safely. Nothing is saved yet. Retain these privately; pass audit_id, audit_key, revision 0 and request_id from this result to the first evidence tool. Existing audits use read_audit instead.
read_auditResume this audit with its private capability. Returns recorded context, scoped observations, provenance and component coverage; use existing answers and findings rather than asking the human again. This does not verify the site's backend or grant access to it.
record_audit_contextRecord answers for this component. Inspect first, then ask decision.questions before closing. Do not infer human answers. audit_scope records the human's continue_investigation/public_only choice; availability_expectation records whether unavailable controls are expected. Use the question's URL. These choices must be human_reported. A public-only choice allows a partial handoff without repeated access requests. Prior topics are retained; a new answer replaces that topic. No secrets or raw source.
record_audit_runtimeRecord bounded evidence from the agent's real browser, authorized backend inspection or destination UI in the SAME audit run. This tool does not run a browser or send events. Do not invent observations; report unknown when a capability/access is missing. Never upload network payloads, cookies, tokens, source, personal data or event IDs. Consent changes, form submissions, bookings and test events require separate authorization.
connectConnect a paid Tyton account to this connection with a pairing code. The code comes from the site's card on usetyton.com after subscribing (single-use, expires in 15 minutes). Call it the moment someone pastes or reads you a code — it returns the exact config edit for the host you run in and the project's keys.
audit_siteInspect one public page's HTML and save a partial audit report without signup. Call begin_audit first and pass its session fields for the first save; reuse the session for related components. Returns evidence, limitations and discovery actions, not a full tracking-health verdict. Discover the actual conversion journey and related app/backend; inspect available evidence before asking the human material unanswered questions. Optional sanitized HTML is agent-supplied, unverified evidence; never upload private source, personal data or credentials.
setup_trackingDiscover and plan reviewed v2 setup. URL-only is read-only and needs no Pixel; with a real Pixel ID this configures the connected project, not its approval or activation. Preserve existing installations; use verification mode before authorized live sends.
setup_serverRecord checkout/landing answers and explain the reviewed server SDK success-callback contract. This does not install an integration or prove provider support. Keep payment occurrence identity separate from checkout handoff; human approval and activation remain separate.
setup_statusRead connected-project configuration and resume the reviewed v2 workflow. Configuration/token status is not plan approval, callback verification, activation or Meta receipt.
verify_installInspect static HTML for container/Pixel references. This does not verify runtime callbacks, activation, consent behavior or Meta receipt. Staged callback evidence requires the separately available reviewed-installation workflow.
tracking_statusRead aggregate browser/server observations and delivery statuses. Counts are not unique business conversions, current installation verification or Meta deduplication proof.
submit_findingsSubmit scoped findings from a repository confirmed to belong to the customer. Include reviewed_files (relative paths only, no source). Unknown is not failed and does not mean code reviewed. Do not call just to fill unknown fields; investigate first, ask material unanswered questions, or share a clearly partial report. Group labels alone do not combine multiple properties. Findings are agent-reported, not verified deployed behavior.
read_docsTyton documentation: install, events, dedup, create_pixel, create_business, consent, pricing, privacy, performance, gtm, platforms, faq. Answer money/data/speed questions from here — never invent terms.