MCP YouTube Transcript
An MCP server retrieving transcripts of YouTube videos
Descripción del creador (original en inglés)
Self-driving LinkedIn prospecting: it sources, warms and invites daily. You approve every message.
Estos tres nunca se combinan en una sola puntuación: las opiniones, las mediciones y la popularidad son cosas distintas.
Todavía no escribimos un resumen sencillo para este. Abajo están los datos que leímos del registro.
Elige la app de IA que usas y copia la configuración.
AuthorizationLos valores marcados como secretos son llaves de tu cuenta. Nunca los compartas ni los pegues en un chat.
Necesitas
¿Qué app usas?
Elige tu app arriba. La recordamos, así que la próxima vez el botón dirá “Instalar en …”.
Esta dirección necesita una clave en un header de la solicitud, y el menú para agregar conectores no puede ponerla. Usa Claude Code, Cursor o VS Code.
Ejecuta esto en tu terminal
claude mcp add --transport http linkedin https://agent.gtmotto.com/mcp --header 'Authorization: your-value'Completa los valores vacíos antes de ejecutarlo.
Pégalo en .cursor/mcp.json de tu proyecto, o en ~/.cursor/mcp.json
{
"mcpServers": {
"linkedin": {
"url": "https://agent.gtmotto.com/mcp",
"headers": {
"Authorization": "your-value"
}
}
}
}Tu navegador te pide permiso para abrir la app; luego la app muestra la configuración y te pide confirmar. Los valores de las claves quedan vacíos para que los completes tú.
Pégalo en .vscode/mcp.json de tu proyecto
{
"servers": {
"linkedin": {
"type": "http",
"url": "https://agent.gtmotto.com/mcp",
"headers": {
"Authorization": "your-value"
}
}
}
}Tu navegador te pide permiso para abrir la app; luego la app muestra la configuración y te pide confirmar. Los valores de las claves quedan vacíos para que los completes tú.
Esta dirección necesita una clave en un header de la solicitud, y el menú para agregar conectores no puede ponerla. Usa Claude Code, Cursor o VS Code.
Tu elección se guarda en una cookie en este dispositivo; no necesitas cuenta.
Authorization secreto HTTP headerEsta configuración se genera a partir de la ficha del creador en el registro. Si el README dice otra cosa, sigue el README.
Todavía no hemos leído la lista de herramientas de este servidor; revisamos los endpoints por turnos.
Revisamos información que cualquiera puede ver; no hemos leído el código. Las revisiones que no aplican no se cuentan. · Revisado el 7 nov 2026
El nombre en el registro se verificó con el dominio gtmotto.com (el registro comprueba quién es dueño del dominio)
Todavía no hemos leído este repo en GitHub
Todavía no hemos leído este repo en GitHub
Aún sin revisar; revisamos los endpoints por turnos
Aún sin revisar; revisamos los endpoints por turnos
Aún sin revisar; leemos la lista de herramientas cuando revisamos el endpoint
Comparado con todos los servidores activos de esta categoría en el registro. Las estrellas solo cuentan cuando quien lo publica es dueño del repo.
Superó 1 de 1 revisiones básicas aplicables. No hemos leído el código. Antes de agregar una clave o un token, revisa qué acceso pide el servidor y usa una clave que tenga solo los permisos que necesitas.
Debes completar: Authorization
Se ejecuta en la nube del creador. Te conectas por internet sin instalar nada, pero lo que envías pasa por los servidores del creador.
El repo de GitHub cambió por última vez –, y la última versión (1.2.1) llegó al registro el mes pasado.
Las apps compatibles con MCP, como Claude Desktop, Claude Code, Cursor y VS Code. Se conecta por URL, así que también funcionan las apps que aceptan servidores MCP remotos, como ChatGPT en modo desarrollador.
En la misma categoría: MCP YouTube Transcript, Google Analytics MCP, BibiGPT YouTube Video Summarizer
Servidores de la misma categoría hechos por otros creadores, ordenados por estrellas.
Cómo leer una tarjeta:★ Estrellas de GitHubiTres pilares: reseñas · revisiones · usoiActualizado = fecha en que se publicó la última versión en el registroi
An MCP server retrieving transcripts of YouTube videos
GA4 MCP server for AI agents: schema discovery, server-side aggregation, analysis-ready defaults.
AI video summarizer: summarize and transcribe YouTube, Bilibili, TikTok and podcast videos.
Copiado tal cual del repo del creador. PopNx MCP no lo ha revisado ni lo respalda. Se quitaron HTML, imágenes y scripts. Los enlaces llevan fuera de este sitio.
Give Claude a LinkedIn seat that works while you sleep.
GTM Otto is self-driving LinkedIn prospecting. From Claude, Cursor or ChatGPT you describe who you want to meet; GTM Otto finds them daily, warms them up from your own LinkedIn account (visit, like, comment), invites them, and opens the conversation. Nothing is said without you.
This repository is the public home of the GTM Otto MCP server: the registry manifest (server.json), the Claude Code plugin with eight skills, the tool reference and the onboarding guide. The server itself runs at https://agent.gtmotto.com/mcp (Streamable HTTP).
Release status (2026-09-22): listed on the official MCP Registry as com.gtmotto/linkedin (manifest 1.2.1). https://agent.gtmotto.com/mcp answers initialize with no credentials; OAuth is served at /.well-known/oauth-protected-resource (authorization server clerk.gtmotto.com); 21 tools are deployed, including gtm_connect_linkedin and the three approval tools.
Every other LinkedIn MCP is a remote control for a tool you still have to operate: lists to upload, sequences to build, seats to rotate. GTM Otto is the only one where the outreach keeps running after the chat ends. A play is an ICP plus up to five sourcing lanes plus an autonomy level. The server exposes 21 tools and, deliberately, no send verb: the agent can set up, size, preview, launch, pause and review a play, and it can approve or reject one drafted message at a time, but it cannot write to a stranger on your behalf.
The endpoint is the same everywhere: https://agent.gtmotto.com/mcp.
Authentication. Add the server with no credentials. The first time the agent calls a protected tool, the client shows a Connect card; sign in (or sign up) in the popup and the call is retried. Scripts and clients that cannot do OAuth can send an API key instead: Authorization: Bearer gtm_sk_..., minted at app.gtmotto.com/connect.
Settings, then Connectors, then Add custom connector. Name it GTM Otto, paste https://agent.gtmotto.com/mcp, save. Start a chat and say what you want; the Connect card appears on the first protected call.
claude mcp add --transport http gtmotto https://agent.gtmotto.com/mcp
Then /mcp inside a session to authenticate in the browser. To use an API key instead:
claude mcp add --transport http gtmotto https://agent.gtmotto.com/mcp \
--header "Authorization: Bearer gtm_sk_..."
The deeplink is cursor://anysphere.cursor-deeplink/mcp/install?name=gtmotto&config=eyJ1cmwiOiJodHRwczovL2FnZW50Lmd0bW90dG8uY29tL21jcCJ9, where the config value is the base64 of {"url":"https://agent.gtmotto.com/mcp"}. By hand, add this to .cursor/mcp.json:
{
"mcpServers": { "gtmotto": { "url": "https://agent.gtmotto.com/mcp" } } }
Settings, then Connectors, then Create (developer mode must be on). Name GTM Otto, URL https://agent.gtmotto.com/mcp, authentication OAuth. Enable the connector in a chat from the tools menu.
Any MCP client that speaks Streamable HTTP works: point it at https://agent.gtmotto.com/mcp, with OAuth or the Authorization header above. n8n, Windsurf, VS Code and the reference mcp-remote shim all fit.
The gtm-otto plugin connects the server and adds eight skills that turn a sentence into a running play:
/plugin marketplace add gtmotto/mcp
/plugin install gtm-otto@gtmotto
| Skill | What it does |
|---|---|
launch-a-play | Estimate the ICP, create a draft play, connect LinkedIn, preview each lane, confirm, activate, explain the next 48 hours. |
icp-to-play | A sentence, a URL or a persona into the seven ICP facets, a company count, and a people_search lane on a new draft or an existing play. |
post-discovery | Turn "people complaining about X on LinkedIn" into a Post discovery lane, with a worked French example. |
hiring-signals | "Companies hiring for X" into a Job offers lane that reaches the buyer inside each hiring company, not the recruiter. |
weekly-play-review | Stats, run history and the approval queue folded into a one-page review with concrete changes. |
instantly-to-linkedin | With the Instantly MCP also connected: companies that opened or replied to a campaign become an account list on LinkedIn. |
heyreach-migration | From an exported HeyReach lead list: companies to an account list, facets to the ICP, and an honest note on what does not carry over. |
linkedin-safety-check | Seat state, daily cap and working hours turned into a plain-English health report. |
Each skill names the exact tools and arguments it calls, what to show you, and where it stops to ask.
Full reference with arguments: docs/tools.md.
| Tool | Scope | One line |
|---|---|---|
gtm_whoami | read | Org, plan, lead meter, and the next step to follow. Start here. |
gtm_list_plays | read | Every play with product, status, seat and pipeline spread. |
gtm_show_play | read | One play in full: ICP, lanes, autonomy ladder, cap, working hours, seat. |
gtm_play_stats | read | The pipeline over four windows, each with its delta versus the window before. |
gtm_linkedin_status | read | The play's seat state and today's headroom against LinkedIn's limits. |
gtm_sources_reference | read | The reference for every sourcing lane. Read before configuring one. |
gtm_show_sources | read | All five lanes of a play: config, switch, blockers, last run, memory. |
gtm_source_runs | read | Run history of one lane, newest first. |
gtm_list_approvals | read | Drafted messages waiting for a human, with their approval ids. |
gtm_estimate_icp | write | "About N companies match" for a candidate ICP. Writes nothing. |
gtm_create_play | write | Create a fully configured play. Draft by default; activate flag. |
gtm_update_icp | write | Change a play's seven ICP facets. Only the ones you name. |
gtm_update_play | write | Name, goal, voice, paused, status active or draft, dailyCap, workingHours. |
gtm_configure_source | write | Write a lane's settings (merge). Can switch it on in the same call. |
gtm_set_source | write | Switch a lane on or off. Refused when there is nothing to run on. |
gtm_preview_source | write | Dry-run one lane: size and a 10-person sample. No side effects. |
gtm_run_source | write | Run one lane now. Real leads, real budget. Only when asked. |
gtm_forget_harvested_posts | write | Clear a post lane's memory so a settings change can be re-tested. Costs. |
gtm_connect_linkedin | write | Returns a browser link to connect the LinkedIn seat; then poll gtm_linkedin_status. |
gtm_approve | write | Approve one drafted message by approval id. The client asks you first. |
gtm_reject | write | Reject one drafted message by approval id. The client asks you first. |
gtm_preview_source is free of side effects; gtm_run_source spends real budget and is only called when you asked.account_list, at most 100) and post URLs, never a CSV of individual LinkedIn profiles.server.json registry manifest (com.gtmotto/linkedin)
.claude-plugin/marketplace.json Claude Code marketplace "gtmotto" plugins/gtm-otto/ the plugin: MCP server + eight skills .claude-plugin/plugin.json skills/ /SKILL.md docs/tools.md the 21 tools with arguments docs/onboarding.md the ideal first conversation CHANGELOG.md
com.gtmotto/linkedinThe playbooks that need nothing but LinkedIn (find buyers in comments, ICP to a Sales Navigator search, hiring-signal prospecting by hand, LinkedIn limits, the comment and invite writer) live in gtmotto/gtm-skills. That repository registers no server and needs no account; each of its skills ends with a pointer to the skill here that runs the same recipe daily.
This repository is mirrored out of the GTM Otto monorepo. Issues and pull requests are welcome here and are read; fixes are applied upstream and land in the next mirror, so a PR may be closed with a "merged upstream" note rather than merged in place. Keep prose free of em dashes, and never describe the product as an automated sales rep of any kind: it is self-driving LinkedIn prospecting.
MIT. See LICENSE.