Jam

Load a browser bug report — console, network, events, video — into your coding agent.

Add to Agentman Studio

Opens Agent Studio, where connecting is one click. The connector URL below works in any MCP client.

Paste it into any MCP client. Setup docs

Resources

What the Jam connector does

The Jam MCP server hands a browser bug report to Claude, ChatGPT or any MCP-compatible coding agent. Paste a Jam link and the agent reads that recording's console logs, network requests, user events, screenshots and video. Sign-in is OAuth or a personal access token, and Jam's own tool list reaches further than reading — it comments, files and deletes.

Verified connector

Listed by Anthropic as a partner connector in its Connectors Directory.

Connection checked by Agentman on .

Anthropic states this reflects the level of review a connector received, not a security audit.

Jam tools (6)

  • getDetails
  • getConsoleLogs
  • getNetworkRequests
  • getScreenshot
  • getUserEvents
  • analyzeVideo

Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.

Limits

  • The directory permissions label is wrong in the risky direction. Read sits over a documented surface containing three delete tools. Review the vendor's tool table, not the listing.
  • We could not verify tool behaviour ourselves. The endpoint returns 401 to an anonymous handshake, so no safety annotations were readable. Read/write classification here is Jam's.
  • Prompts are unconfirmed. Anthropic's snapshot lists no prompt names, but the gated endpoint blocked our own prompts/list, so we cannot say whether the server serves any.
  • Redaction covers headers and JSON request bodies. Jam documents no equivalent claim for response bodies or for secrets carried in URL query strings.
  • Auto-blur needs the snippet. It applies to Jams requested through Recording Links or Intercom on sites with a verified Jam.js install, and Jam lists iframes, page-load flashes and window resizes as cases where unblurred content can slip through.
  • Instant Replay Jams are a reduced surface. Jam states getScreenshot and analyzeVideo do not work on them; the other tools do.
  • getFrames requires Cloudflare Stream hosting. Screenshot Jams and videos not yet on Cloudflare return an explanatory text block instead of images.
  • Network bodies are sometimes absent by design. Jam names three cases: the request predated Jam initialising on the page, a CORS preflight failed, or the asset was static rather than XHR, fetch or WebSocket.
  • Deletion is close to irreversible. Jam says a deleted Jam leaves lists and search with no way to restore it, and that a deleted folder takes every Jam inside it.
  • Roles are a paid feature. Jam documents Viewer, Creator and Admin seats as available on Team and Enterprise plans, and puts SSO, audit logs, access controls and custom retention on Enterprise.
  • HIPAA is not available. Jam's security page lists SOC 2 Type II and GDPR as compliant and HIPAA as coming soon, fetched 2026-08-19.

Frequently asked questions

Does the Jam MCP server only read data, or can it write and delete?

It writes and deletes. Jam's MCP documentation lists tools that create comments, rename and move Jams, create folders, and issue recording links — plus `deleteJam`, `deleteComment` and `deleteFolder`. Jam states a deleted Jam cannot be restored and that deleting a folder takes every Jam inside it. Anthropic's directory labels the connector Read, which understates the surface.

Does Jam redact secrets from the network requests an agent can read?

Yes, and it happens before upload. Jam's DevTools page says it scans network requests for tokens, cookies and PII, filters those fields client-side, and substitutes the string `JAM_DOES_NOT_SAVE_SECRETS`. Jam states the filtering runs prior to Jam creation, so the values never reach its servers. Jam also says it prefers to over-filter rather than under-filter.

What does the analyzeVideo tool actually send, and to whom?

It analyses a stored Jam video recording and returns structured findings about what went wrong. Jam's MCP FAQ states that some MCP tools use Google's Gemini, and its security page says Jam opts out of third-party model training, limits vendor retention, and de-identifies content first. The video itself already lives on Jam's infrastructure before any agent asks for it.

Do I need the Jam Chrome extension to use the MCP server?

No. Jam's documentation says you only need an existing Jam link, and that anyone — you, a teammate, or a customer — can be the one who recorded it. The extension or the iOS app is required only if you want to create recordings yourself. The agent consumes a link that already exists in your workspace.

Can an agent see Jams that I cannot see myself?

No. Jam's documentation states that MCP mirrors your existing Jam permissions and grants no access beyond what a user can already reach in the Jam web or mobile apps. Workspace membership, roles and SSO remain the controls. A personal access token is additionally scoped to one workspace and tied to your own user account.

How do I connect Jam MCP in a CI pipeline with no browser?

Use a personal access token instead of the OAuth flow. Jam's documentation describes creating one under Settings, MCP, choosing an expiry of 7, 30 or 90 days or 1 year, and selecting the `mcp:read` or `mcp:write` scope. Tokens carry a `jam_pat_` prefix so secret scanners recognise them, and Jam stores only a hash.

Why does getScreenshot return nothing for my Jam?

It is probably an Instant Replay Jam. Jam's documentation states that `getScreenshot` and `analyzeVideo` are unavailable for Instant Replay captures, while every other MCP tool still returns data. Separately, `getFrames` works only on video Jams hosted on Cloudflare Stream; other recordings get a short text explanation in place of images.

Sources

  • Jam MCP documentation — https://jam.dev/docs/jam-mcp (fetched 2026-08-19; the directory's documentation URL jam.dev/docs/debug-a-jam/mcp returns 308 to this page) · retrieved 2026-08-19
  • Jam DevTools — https://jam.dev/docs/devtools (fetched 2026-08-19 as Markdown via the .md suffix) · retrieved 2026-08-19
  • Jam Auto-blur — https://jam.dev/docs/auto-blur (fetched 2026-08-19) · retrieved 2026-08-19
  • Jam Security — https://jam.dev/docs/security (fetched 2026-08-19) · retrieved 2026-08-19
  • Jam AI policy — https://jam.dev/docs/ai-policy (fetched 2026-08-19) · retrieved 2026-08-19
  • Jam personal access tokens — https://jam.dev/docs/personal-access-tokens (fetched 2026-08-19) · retrieved 2026-08-19
  • Jam members and roles — https://jam.dev/docs/members-and-roles (fetched 2026-08-19) · retrieved 2026-08-19
  • Jam docs machine index — https://jam.dev/docs/llms.txt (fetched 2026-08-19) · retrieved 2026-08-19
  • jam.dev/docs/robots.txt allows the docs path and carries Content-Signal: ai-train=yes, search=yes, ai-input=yes — synthesis is expressly permitted on all three axes (fetched 2026-08-19). The marketing host jam.dev/robots.txt is a bare Allow: / · retrieved 2026-08-19
  • Live RFC 9728 protected-resource descriptor — https://mcp.jam.dev/.well-known/oauth-protected-resource (fetched 2026-08-19; the path-append and path-insert forms both 404) · retrieved 2026-08-19
  • Live RFC 8414 authorization-server metadata — https://mcp.jam.dev/.well-known/oauth-authorization-server and https://api.jam.dev/.well-known/oauth-authorization-server (both fetched 2026-08-19; the issuer's copy advertises two additional CLI scopes) · retrieved 2026-08-19
  • Live anonymous probe — 2026-08-19: POST to the endpoint returns HTTP 401 with www-authenticate: Bearer resource_metadata="…", scope="mcp:read mcp:write" · retrieved 2026-08-19
  • Anthropic connector directory — https://claude.ai/directory/923fb14b-ed65-4e41-8dd5-4c2c38c00457 (snapshot 2026-08-16) · retrieved 2026-08-16

More Developer Tools connectors

  • Base44

    Build and edit Base44 apps from an AI assistant, and query data your app's users submit.

  • CircleCI

    Diagnose failed builds, read job logs, inspect test results and rerun or roll back deployments from your AI assistant.

  • Claude Managed Agents

    Create, run, monitor and debug Anthropic's Claude Managed Agents from inside Claude itself.

Browse all MCP connectorsClaude and ChatGPT setup guideBrowse AI skills

Product names and logos belong to their owners. Agentman is not affiliated with a connector's maker unless the page says the connector is built and run by Agentman.

Use it in an agent

Put Jam to work.

Connect Jam once and your agents call these tools on their own: on a schedule, in a workflow, with nobody at the keyboard.

Open in Agentman Studio