For developers & agents

Your podcast intelligence, in any agent.

Connect Claude Code, claude.ai, or any MCP client to Vuci over a single authenticated endpoint — and ask questions of the index in plain language. Search transcripts, pull episode and podcast metadata, topics, your Spaces and Alerts. Read-only, token-authenticated, grounded in real audio.

Get access See the tools ↓

Read-only · self-managed tokens · works with any MCP client

Connect https://vuci.ai/mcp/

# add Vuci to Claude Code

claude mcp add --transport http vuci

https://vuci.ai/mcp/ \

--header "Authorization: Bearer vuci_live_…"

→ 8 read-only tools registered

search_episodes · get_episode · get_topic …

your data, one question away

Get started

Connected in 30 seconds.

Get on a plan with access

MCP access is included with the Podcaster and Enterprise plans.

Generate a token

Mint a token from your account settings. It's shown once — copy it then; we only ever store a hash.

Point your client at Vuci

Add https://vuci.ai/mcp/ as an HTTP MCP server with your token as a Bearer credential — done.

Claude Code

claude mcp add --transport http vuci https://vuci.ai/mcp/ \ --header "Authorization: Bearer vuci_live_xxxxxxxx_..."

Using claude.ai or another MCP client? Add a custom HTTP MCP server with the URL https://vuci.ai/mcp/ and your vuci_live_… token as a Bearer credential.

The tools

11 read-only tools, one endpoint.

Every tool reads from Vuci's transcript index and returns structured, source-linked data. None can mutate your account.

Search

  • search_episodes

    Search Vuci's indexed podcast content (podbits, quotes, takeaways, descriptions) using keywords.

    Parameters

    query string required Search query (2–200 characters)
    podcasts array optional Optional list of podcast slugs to filter by
    categories array optional Optional category slugs to filter by
    types array optional Content types to search
    date_from string optional ISO date YYYY-MM-DD lower bound
    date_to string optional ISO date YYYY-MM-DD upper bound
    sort string optional Sort order (default: relevance) (relevance, date_desc, date_asc, podcast_asc)
    page integer optional Page number (default: 1)
    page_size integer optional Results per page, max 20 (default: 10)

Content

  • get_episode

    Retrieve full metadata for a single podcast episode by its slug.

    Parameters

    slug string required Episode slug

  • get_podcast

    Retrieve metadata for a single podcast by its slug.

    Parameters

    slug string required Podcast slug

  • podcast_analytics

    Return analytics for a podcast: publishing activity, duration stats, frequent guests, content DNA.

    Parameters

    slug string required Podcast slug

  • get_topic

    Retrieve a published topic by its slug.

    Parameters

    slug string required Topic slug

Workspace

  • list_spaces

    List all spaces owned by the authenticated user.

    No parameters

  • get_space

    Retrieve a single space by ID. Only returns your own spaces or public spaces.

    Parameters

    id integer required Space primary key

  • get_space_items

    List the items saved inside a space (episodes, podbits, podcasts, quotes, snapshots, topics, notes). Only your own spaces or public spaces. Supports paging and a content_type filter.

    Parameters

    id integer required Space primary key
    content_type string optional Optional: only items of this kind (podbit, episode, podcast, quote, takeaway, topic, note, url)
    page integer optional Page number (default 1)
    page_size integer optional Results per page, max 50 (default 20)

  • list_alerts

    List all alerts (notification rules) owned by the authenticated user.

    No parameters

  • get_alert

    Retrieve one alert (notification rule) by ID, including its keyword, company, podcast, topic and person targets and its delivery settings.

    Parameters

    id integer required Alert (notification rule) primary key

  • list_alert_notifications

    List fired notifications across your alerts (newest first), optionally filtered to a single alert. Each notification links the matched episode, trigger and transcript snippet.

    Parameters

    alert_id integer optional Only notifications from this alert
    page integer optional Page number (default 1)
    page_size integer optional Results per page, max 50 (default 20)

What it's grounded in

Real moments, pulled from real audio.

Your agent reads the same index the rest of Vuci is built on — the podcasts it covers, clustered into the topics they keep returning to.

TopicBottled water contains an average of 240,000 microplastic…Trending now TopicClinical rates of depression have tripled and anxiety has…Trending now TopicAircraft Engine Lease Rate SofteningTrending now

Safe by design

Connecting an agent can't break things.

Read-only scope

Tokens carry a read scope only. A connected agent can query your data — it can never create, edit, or delete anything in your account.

Self-managed tokens

Mint and name tokens yourself, and revoke any of them instantly from your settings — no support ticket, effective immediately.

Hashed secrets

The full token is shown once at creation and never again — we store only a hash, so a database read can't reconstruct your credential.

Per-token rate limiting

Each token is rate-limited independently to keep usage fair and the service stable for everyone.

Podcaster or Enterprise plan required

Connect Claude and any MCP-compatible agent to your Vuci podcast intelligence — search, episode insights, topics, Spaces, and Alerts. Requires a Podcaster or Enterprise plan.
  • Read-only MCP server at vuci.ai/mcp
  • Works with Claude Code, claude.ai, and any MCP client
  • Self-managed API tokens with instant revoke
View plans

FAQ

Questions, answered.

Does Vuci have an MCP server?

Yes. Vuci runs a remote MCP (Model Context Protocol) server at https://vuci.ai/mcp/ that lets Claude Code, claude.ai, or any MCP-compatible client query its podcast intelligence — searching transcripts and pulling episode, podcast, topic, Space, and Alert data — over an authenticated, read-only connection.

How do I connect Claude to Vuci?

Generate a token at vuci.ai/settings/mcp/, then run 'claude mcp add --transport http vuci https://vuci.ai/mcp/ --header "Authorization: Bearer <your-token>"'. For claude.ai or any generic MCP client, add a custom HTTP MCP server with the URL https://vuci.ai/mcp/ and your token as a Bearer credential.

What can the Vuci MCP server do?

It exposes read-only tools to search indexed podcast content; fetch a single episode, podcast, or published topic; pull podcast analytics; list the caller's own Spaces and read the items inside them; and list the caller's Alerts, one Alert's detail, and the notifications those Alerts have fired. Every tool returns structured, source-linked data grounded in Vuci's transcript index.

Can an agent change my Vuci account through the MCP server?

No. Every tool is read-only — the issued token carries a read scope only, so a connected agent can query your data but cannot create, edit, or delete anything. Tokens are self-managed and can be revoked instantly.

Who can use the Vuci MCP server?

MCP access is included with the Podcaster and Enterprise plans. Anyone can read these docs; minting a token requires an entitled plan, after which tokens are managed from your account settings.

Is the connection secure and rate-limited?

Yes. Token secrets are stored hashed (shown once at creation, never again), each request is authenticated by Bearer token, and the endpoint is rate-limited per token to keep usage fair and the service stable.