Click on "Install 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., "@Second Brain MCP What did the Lex Fridman podcast say about AGI timelines?"
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.
Remember everything you watch and listen to.
A Model Context Protocol server that turns YouTube videos and podcasts into a private, searchable memory — built on the brand-new stateless MCP spec (2026-07-28).
Paste a link once. Ask about it forever.
"What did that video I watched last month say about salary negotiation?"
🧠 → "At [12:43] in Never Split the Difference — Chris Voss, he says never to accept the first offer without anchoring high…" — with a deep link that jumps to the exact second.
You watch hours of talks, tutorials, and podcasts — and a week later you can quote none of it. Browser history remembers that you watched something; nothing remembers what it said. Second Brain gives your AI assistant total recall over everything you've ever watched or listened to:
🎯 Ask across your whole watch history — answers come back with timestamps and deep links to the exact moment.
🔒 Private by design — one local SQLite file. No cloud, no accounts, no API keys. Podcast audio is transcribed locally.
⚡ Zero-key ingestion — YouTube captions are fetched directly; nothing to configure.
🗑️ Nothing is deleted without you — destructive operations use the spec's new multi round-trip (MRTR) approval flow.
Related MCP server: yt
Option A — as a tool, straight from GitHub (recommended):
uv tool install git+https://github.com/ravishu5/second-brain-mcpThis installs the second-brain command into ~/.local/bin. Find its absolute path — you'll need it below:
which second-brain # e.g. /Users/you/.local/bin/second-brainOption B — from a local clone (for hacking on it):
git clone https://github.com/ravishu5/second-brain-mcp
cd second-brain-mcp
python3 -m venv .venv && .venv/bin/pip install -e .Your binary is then at <clone-dir>/.venv/bin/second-brain.
Always use the absolute path to the binary. GUI apps (like Claude Desktop) don't inherit your shell's
PATH, so a baresecond-braincommand often fails with "Failed to spawn process: No such file or directory".
Claude Code
claude mcp add second-brain -- /absolute/path/to/second-brainClaude Desktop — Settings → Developer → Edit Config, then add:
{
"mcpServers": {
"second-brain": {
"command": "/absolute/path/to/second-brain"
}
}
}
ex: "command": "/Users/ravi/Desktop/mcp/second-brain-mcp/.venv/bin/second-brain" in my caseFully quit (⌘Q) and reopen Claude Desktop — the server should show as connected under Settings → Developer.
As a stateless HTTP service (deployable behind any load balancer — no sticky sessions, no shared state):
second-brain --http --port 8000Talk to your assistant:
› Remember this: https://www.youtube.com/watch?v=8S0FDjFBj8o
› What have I saved about system design?
› What did Lex's guest say about AGI timelines? Link me to the moment.
› Give me a digest of everything I added this week.Tool | What it does |
| Ingest a YouTube video or podcast episode: fetch transcript, chunk with timestamps, index. |
| BM25 full-text search across every transcript; returns passages with deep links to the exact second. |
| Everything in the brain, newest first, with stats. |
| Read a raw transcript, optionally sliced by time range. |
| What you added recently, rolled up. |
| Delete an item — gated by MRTR user confirmation. |
Podcast transcription is optional (local faster-whisper):
pip install "second-brain-mcp[whisper]"This server is a working showcase of MCP's biggest release since remote MCP launched:
sequenceDiagram
participant U as User
participant C as Client (Claude)
participant S as Second Brain
C->>S: tools/call forget("that crypto video")
S-->>C: resultType: input_required ("Permanently delete? No undo.")
C->>U: Asks for approval
U-->>C: Yes
C->>S: tools/call retry (inputResponses + requestState)
S-->>C: "Forgot 'Crypto Explained' (213 chunks removed)."Stateless by construction. No initialize handshake, no Mcp-Session-Id. Every request is self-contained; protocol metadata rides in _meta per request. Run one instance or twenty behind a load balancer — nothing breaks, because the only durable state is your local SQLite file, addressed through explicit item ids that clients thread between calls (exactly the application-state pattern the spec prescribes).
MRTR instead of server-push. forget returns resultType: "input_required" with an elicitation request; the client gathers your approval and retries with inputResponses. No open streams, no sessions — and nothing is ever deleted without a human saying yes.
Header-routable. Under the streamable HTTP transport, gateways can rate-limit Mcp-Method: tools/call + Mcp-Name: remember (expensive ingestion) differently from cheap recall reads — without parsing a byte of JSON.
SDK v2. Built on mcp v2.0.0, released alongside the spec. Type hints are the schema; Resolve() powers the MRTR flow.
┌─────────────┐ remember(url) ┌──────────────┐
│ Claude / │ ────────────────► │ ingest.py │ YouTube captions (no key)
│ any MCP │ │ │ RSS + local whisper
│ client │ recall(query) ├──────────────┤
│ │ ────────────────► │ store.py │ SQLite + FTS5 (BM25)
└─────────────┘ ◄─ passages + │ ~/.second- │ timestamped chunks
deep links │ brain/ │
└──────────────┘Transcripts are merged into ~450-character chunks, split on silence gaps (usually topic boundaries), each carrying its start/end timestamps — so search hits map back to a playable moment, not just a document.
"Server disconnected" / "Failed to spawn process: No such file or directory" — the client can't find the binary. Use the absolute path from which second-brain (or your venv's .venv/bin/second-brain) in the config, then fully restart the client.
macOS: PermissionError: Operation not permitted on a path under ~/Desktop, ~/Documents, or ~/Downloads — macOS blocks other apps from reading those folders. Don't point the client at a venv inside them; install outside instead: uv tool install git+https://github.com/ravishu5/second-brain-mcp (lands in ~/.local, which isn't protected). Alternatively grant the client Desktop access under System Settings → Privacy & Security → Files and Folders.
"No transcript/captions available" — the video has captions disabled; there's nothing to index (yet — see roadmap).
Podcast ingestion errors about whisper — install the optional extra: pip install "second-brain-mcp[whisper]".
Where's my data? One SQLite file at ~/.second-brain/brain.db. Override with the SECOND_BRAIN_DB env var. Delete the file to wipe the brain.
git clone https://github.com/ravishu5/second-brain-mcp
cd second-brain-mcp
uv sync --extra dev
uv run pytest
uv run mcp dev src/second_brain_mcp/server.py # MCP inspectorClient-side summaries on ingest via MRTR sampling/createMessage (the client's model writes the summary — still zero server keys)
Semantic search (local embeddings) alongside BM25
Browser extension: one-click "remember this"
Whisper speaker diarization for podcasts
PRs welcome — especially ingestion sources (lectures, audiobooks, Twitch VODs).
MIT © Ravi Shankar
Summarize what was added to the brain in the last N days.
| Name | Required | Description | Default |
|---|---|---|---|
| days | No |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description must disclose behavioral traits. It states the core action but does not mention whether it is read-only, what the summary contains, how the 'N days' parameter behaves, or any other side effects. It provides minimal additional context beyond what the name and schema suggest.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single sentence with no unnecessary words. It is front-loaded with the action and resource, making it immediately understandable.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
This is a simple tool with one optional parameter and an output schema, so the description need not explain return values. It covers the core functionality. However, it lacks any context about what 'the brain' refers to or how the summary is structured, which would be helpful but not essential given the tool's simplicity.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema provides only 'days' with a default, and the description clarifies that it represents 'the last N days'. This adds meaning beyond the bare parameter name. Given the schema description coverage is 0% and there is a single parameter, the description compensates reasonably well by linking the parameter to the time window.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description uses a specific verb ('Summarize') and a clear resource ('what was added to the brain in the last N days'). This distinguishes it from sibling tools like 'recall' (which likely retrieves specific information) and 'forget' (which removes).
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage: when you need a summary of recent additions over a time window. However, it does not explicitly state when to use this tool versus alternatives, nor does it mention any exclusions or prerequisites.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Permanently delete an item (by id, URL, or title fragment).
Requires explicit user confirmation via the MRTR elicitation flow — nothing is deleted until the user approves.
| Name | Required | Description | Default |
|---|---|---|---|
| item_ref | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It adequately discloses the destructive nature ('Permanently delete') and the confirmation requirement ('nothing is deleted until the user approves'), which are key behavioral traits. It does not cover side effects or failure modes, but these are not critical for this simple tool.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is two sentences long, front-loaded with the primary action, and each sentence adds essential information: the deletion action and the confirmation requirement. There is no redundancy or filler.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple one-parameter tool with an output schema, the description covers all essential aspects: what it does, what it accepts, and the critical confirmation flow. No missing information is needed for correct usage.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema only names the parameter 'item_ref' with no description, and schema coverage is 0%. The description compensates by explaining that the reference can be an id, URL, or title fragment, giving meaningful semantics beyond the schema. It could be more specific about the format of the id or fragment, but it provides sufficient guidance.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool's action ('Permanently delete an item') and specifies the target resource ('by id, URL, or title fragment'). It is distinct from sibling tools like 'remember' and 'recall', establishing a unique purpose for deletion.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides clear context on when to use the tool (for deletion) and includes an important prerequisite (explicit user confirmation via MRTR elicitation flow). However, it does not explicitly mention alternatives or when not to use the tool, so it falls short of a 5.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
List everything stored in the second brain, newest first.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The description discloses 'newest first' as a sorting behavior but omits the fact that the 'limit' parameter (default 25) constrains the result set. This is a significant behavioral omission that could mislead an agent into thinking all stored items are returned when actually only a limited subset is.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is one concise sentence that efficiently states the core action and ordering. Every word contributes meaning, and it is appropriately front-loaded.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Even though the tool has only one optional parameter and an output schema exists, the description fails to mention the limiting behavior or provide any usage context. The agent cannot fully anticipate the output scope or how to request larger result sets.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema defines 'limit' with a default but provides no description, and the tool description does not mention or explain this parameter. With 0% schema description coverage, the agent cannot accurately understand what 'limit' does or how it affects results.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description uses a specific verb 'List' and identifies the resource as 'everything stored in the second brain,' with ordering 'newest first.' It clearly conveys a comprehensive listing action and implicitly distinguishes from more targeted sibling tools like 'recall' or 'remember.'
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool versus siblings like 'recall' or 'digest.' It lacks explicit alternatives, exclusions, or context for appropriate use, leaving the agent to infer when 'library' is the right choice.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Search everything ever watched or listened to.
Full-text (BM25) search across all indexed transcripts. Returns the most relevant passages with timestamps and deep links so the exact moment can be replayed. Use this to answer questions like "what did that video say about X?" — then answer from the returned passages, citing timestamps.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | ||
| query | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the burden. It discloses search algorithm (BM25), output format (relevant passages with timestamps and deep links), and relevance ordering. As a read-only search tool, this is adequate behavioral disclosure.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise (3 sentences) and front-loaded with the core action. Every sentence adds value, though the first two sentences are somewhat redundant ('Search everything...' and 'Full-text search...') but not bloated.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple search tool with an output schema (not shown), the description explains return values and usage. It lacks details on `limit` semantics and scope (e.g., whether 'everything' is user-specific), but overall it is sufficiently complete for a low-complexity tool.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has 0% description coverage and the description does not directly explain `limit` or define `query` beyond an example. The example implies query is a natural language question, but `limit` is entirely omitted, leaving parameter semantics under-specified.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description starts with a clear verb+resource: 'Search everything ever watched or listened to.' It further specifies full-text BM25 search across indexed transcripts, distinguishing from siblings like 'transcript' (retrieving one transcript) or 'library' (listing).
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides explicit usage context: 'Use this to answer questions like "what did that video say about X?"' and instructs to answer from returned passages with citations. It does not explicitly state when not to use, but the use case is clear.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Add a YouTube video or podcast episode to the second brain.
Pass a YouTube URL, or a podcast RSS feed URL (optionally with an
episode title filter; defaults to the latest episode). The transcript
is fetched, chunked with timestamps, and indexed for search. Returns the
stored item's id — use it with transcript and forget.
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | ||
| episode | No |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the behavioral burden. It discloses that the transcript is fetched, chunked with timestamps, and indexed for search, and that the tool returns the stored item's id. This is meaningful process transparency, though it omits potential edge cases like duplicate URLs or unsupported media types.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is three sentences long, front-loads the core purpose, and each sentence adds distinct value: input types, processing behavior, and output usage. No filler or repetition of schema info.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a tool with two parameters and an output schema, the description covers input, processing pipeline, and return value, and connects to related tools. It lacks explicit constraints like authentication or format limitations, but given the output schema, this is sufficiently complete for an agent to use the tool correctly.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 0%, so the description fully compensates. It explains that `url` accepts a YouTube URL or podcast RSS feed URL, and that `episode` is an optional title filter defaulting to the latest episode. Both parameters are semantically enriched beyond the bare schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a specific action and target: 'Add a YouTube video or podcast episode to the second brain.' It clearly distinguishes this from sibling tools (recall, library, transcript, digest, forget) by focusing on ingestion, not retrieval or deletion.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
It explains how to use the tool: pass a YouTube URL or podcast RSS feed, optionally with an episode filter defaulting to the latest episode. It also says the returned id should be used with `transcript` and `forget`, giving follow-up context. It does not explicitly state when not to use it, but the usage context is clear.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Read the raw transcript of one item (by id, URL, or title fragment).
Optionally slice by start/end seconds to pull just one section.
| Name | Required | Description | Default |
|---|---|---|---|
| end | No | ||
| start | No | ||
| item_ref | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the burden of disclosing behavior. It communicates a read-only operation via 'Read' and adds the slicing behavior with start/end seconds. It does not mention error handling or response size, but the output schema exists and the behavior is simple.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Two concise, front-loaded sentences deliver the core purpose and an optional feature without any wasted words. The structure is easy to scan and process.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple read tool with an output schema, the description adequately covers required input formats, optional parameters, and the slicing use case. It does not address edge cases like multiple title matches or not-found behavior, but these are not critical for a straightforward retrieval tool.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 0%, so the description compensates well. It explains that item_ref can be an id, URL, or title fragment, and that start/end are in seconds and optional for slicing. This adds meaningful semantics beyond the bare schema titles.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a specific verb and resource ('Read the raw transcript of one item') and clarifies acceptable reference types (id, URL, or title fragment). This clearly distinguishes it from sibling tools like 'digest' or 'recall' by focusing on raw transcript retrieval.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides clear context: use when you need the raw transcript of a single item, optionally limited to a time slice. It does not explicitly name alternatives or when-not-to-use conditions, but the purpose is specific enough to imply the right scenario.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
v0.1.0digestforgetlibraryrecallremembertranscriptEach tool serves a unique function: remember adds content, recall searches, library lists, transcript retrieves raw text, digest summarizes recent additions, and forget deletes. No two tools overlap in purpose, and their descriptions clarify the distinct use cases.
Most tool names are single-word imperative verbs (remember, recall, digest, forget), giving a clear action-oriented pattern. However, 'library' and 'transcript' are nouns, breaking the otherwise consistent verb style, though still predictable in context.
With six tools, the set is well-scoped for a personal knowledge management server. Each tool addresses a core need—adding, searching, listing, reading, summarizing, and deleting—without unnecessary bloat or overlap.
The tool surface covers the full lifecycle for stored media: creation (remember), retrieval (recall, library, transcript), summarization (digest), and deletion (forget). There are no obvious gaps or dead ends; the only potential missing operation is update, but it's not essential for this domain.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
An MCP server that provides tools to discover and retrieve podcast episodes transcripts.
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.
MCP server for RiverScript, an AI transcription platform - fetches transcripts shared via a link.
MCP server for structured access to Lenny Rachitsky podcast transcripts. For content creators.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/ravishu5/second-brain-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
