Provides search and retrieval of indexed YouTube and podcast transcripts, including full transcripts, episode listings, mentions, momentum, sponsors, and coverage.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@arcmira Show me Ramp's entity momentum over the last 30 days"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Arcmira MCP is the official remote MCP server for Arcmira. It gives any MCP client ten read-only tools over indexed YouTube and podcast transcripts: the full transcript of one video, the newest episodes of a show, what a show said, who was mentioned where, momentum, sponsors, organic recommendations, and coverage.
Arcmira is an SF-based AI company and the search engine for the spoken web.
claude mcp add --transport http arcmira https://mcp.arcmira.com/mcpOne remote URL. Sign in through your host, or send an account key. The server is a stateless facade over the public HTTP API at https://api.arcmira.com/v1; every gate the API raises is forwarded untouched with the link that lifts it.
Endpoint: https://mcp.arcmira.com/mcp (Streamable HTTP)
Registry name: io.github.arcmira/arcmira
Overview: https://arcmira.com/mcp
Setup for each host: https://arcmira.com/agent-setup
Reference: https://arcmira.com/docs/mcp-server
Server card: https://mcp.arcmira.com/.well-known/mcp/server-card.json
HTTP API the tools front: https://api.arcmira.com/v1/openapi.json
Two ways in. Hosts that speak the MCP authorization spec sign you in; everything else sends a key.
Sign in through the host. Add https://mcp.arcmira.com/mcp with no key. The server answers 401 with an OAuth challenge, the host registers itself against api.arcmira.com, opens arcmira.com for sign-in and consent, and connects with a token that carries the permissions you allowed. Tokens refresh on their own; revoke a host under Settings, Connected apps.
claude mcp add --transport http arcmira https://mcp.arcmira.com/mcpcodex mcp add arcmira --url https://mcp.arcmira.com/mcpClaude Desktop, claude.ai, ChatGPT, and Cursor: add the URL as a custom connector or MCP server with no headers and follow the sign-in prompt. Per-host steps are on https://arcmira.com/agent-setup.
Send a key. Any client that cannot do the sign-in sends a bearer token instead, and the server skips OAuth.
An account key (arc_sk_...) comes from https://arcmira.com. Plan and scopes decide what each tool returns. With no account and no browser, sign up from the API. Post an email address, then post the six digit code from that inbox back.
curl -X POST "https://api.arcmira.com/v1/signups?src=mcp-tool" \
-H 'Content-Type: application/json' \
-d '{"email":"agent@example.com"}'
curl -X POST "https://api.arcmira.com/v1/signups/verify" \
-H 'Content-Type: application/json' \
-d '{"email":"agent@example.com","code":"482913"}'claude mcp add --transport http arcmira https://mcp.arcmira.com/mcp --header "Authorization: Bearer $ARCMIRA_API_KEY"{
"mcpServers": {
"arcmira": {
"url": "https://mcp.arcmira.com/mcp",
"headers": { "Authorization": "Bearer arc_sk_..." }
}
}
}The 401 body carries that signup call under error.data.unlock.action, so an agent that cannot sign in can create an account and reconnect. Discovery: https://mcp.arcmira.com/.well-known/oauth-protected-resource names the authorization server; https://api.arcmira.com/.well-known/oauth-authorization-server lists its endpoints.
Related MCP server: YouTube MCP Server
Tool | Use it for | Fronts |
| Turn a name, |
|
| Short spoken slices for one topic, with watch links and dates |
|
| Has X mentioned Y yet, first seen, last seen |
|
| Mentions in the last 7 and 30 days against the prior 30, with a verdict |
|
| What a set of shows talks about, and what they share; |
|
| The newest indexed episodes of one channel, with the |
|
| Recurring sponsors of a channel from the ad-read rollup |
|
| Who recommends an entity on air, each row |
|
| What the index holds for a channel, or one transcription job |
|
| Full transcript of one video from its URL or id, as text with |
|
Every tool declares readOnlyHint: true, destructiveHint: false, idempotentHint: true, openWorldHint: false. Every result carries as_of where a date applies and a one-sentence note. Nine of the tools return the JSON body in both the content block and structuredContent. get_transcript returns two text blocks and no structuredContent: the transcript as [start] text lines, then the rest of the body as JSON. A long transcript crosses the wire once, and a host that prefers structuredContent still shows the lines.
Good first calls: TBPN is channel UC-DRzaGnL_vtBUpCFH5M0tg, Moment of Truth is UClWkDGXEzsh77GAhs90wpXw, Ramp is ent_14.
A gate is an MCP tool result with isError: true whose content is the API's error envelope:
{
"error": {
"type": "permission_error",
"code": "filter_requires_paid",
"message": "The source filter needs Hobby.",
"param": "source",
"gate": "plan",
"unlock": { "tier": "hobby", "url": "https://arcmira.com/pricing?src=mcp-tool", "offer": null },
"doc_url": "https://arcmira.com/docs/errors#filter_requires_paid",
"request_id": "req_..."
}
}Switch on error.code, relay error.unlock.url to the human, and honor retry_after_seconds on rate_limited. list_recommendations needs a Pro plan; a free key gets recommendations_not_enabled with the unlock link. A 200 that withheld something (Premium transcript text, the paid-versus-organic split, sponsors past the free slice) is a normal result carrying the same body under access. The full catalog is at https://arcmira.com/docs/errors.
Every result, gates included, carries the key's budget after the call under _meta["arcmira.com/rate_limit"] as { "limit": 20, "remaining": 17, "reset": 1788819360 }, read from the API's RateLimit headers. reset is Unix seconds at the next window. Hosts do not show _meta; a client that watches its spend reads it there.
Every result also carries _meta["arcmira.com/build"]: server (this server's version), deploy (its Worker deploy id), api (the deploy id of the API build that answered, from the X-Arcmira-Build response header), and client (the host's name and version from its initialize handshake, which the server forwards to the API as x-arcmira-client). A transcript a host keeps can be joined to the exact code that produced it.
pnpm install
pnpm dev # wrangler dev on :8790, API base from .dev.vars
pnpm test # node:test
pnpm typecheck
pnpm manifest:check # every tool call matches the live OpenAPI document
ARCMIRA_KEY=arc_sk_... node --experimental-strip-types scripts/smoke.ts http://localhost:8790/mcp
ARCMIRA_KEY=arc_sk_... node --experimental-strip-types scripts/measure-transcript-bytes.ts <video-id> http://localhost:8790/mcpThe tool descriptions are loaded context and are the steering surface. They are maintained in Arcmira's manifest spec first and copied here verbatim; revise there before here.
Copyright Arcmira. All rights reserved. See LICENSE.
This server cannot be deployed
YouTube transcripts, search, channel browsing, and playlists for AI agents via MCP.
Search your video library with natural language and retrieve relevant moments through MCP.
YouTube discovery, transcripts, library search, and monitors with API keys or OAuth.
An MCP server that gives any LLM or agent clean YouTube transcripts on demand: a single video, a whole channel, or a playlist, plus AI cleanup of auto-generated captions. API-key auth, credit-based, same backend as the public v1 API. Get a free API key with 25 free credits at youtubetranscriptdownload.com/account.
