Skip to main content
Glama

πŸ§ͺ Distill

CI Python 3.11+ License: MIT

An MCP server that gives Claude Code a shared team knowledge base β€” a local LLM transforms your raw input into anonymous, factual knowledge before anything leaves your device.

No author. No frustration. No names. Just a clean, reusable fact.

Distill demo β€” raw thought to clean team fact

Raw input β†’ local Ollama β†’ review β†’ team DB or discard

Quick Start

pip install distill-mcp
ollama pull gemma3:4b && ollama pull nomic-embed-text
claude mcp add distill -- python -m distill_mcp

Then in Claude Code:

You:    "No, we don't use REST here. We switched to gRPC last month."

Claude: [saves to distill]
        Got it. I've noted that the team uses gRPC, not REST.

        ... next session, different repo ...

You:    "Set up the API for this new service."

Claude: [searches distill β†’ finds gRPC decision]
        Based on your team's knowledge, I'll set up a gRPC
        service since the team switched from REST last month.

Distill saves when you correct a mistake or make a decision, and searches before proposing architecture β€” no prompting needed.

Related MCP server: myBrAIn

What makes this different

Every "memory MCP" stores your raw text in a database. Distill doesn't. The local LLM is a mandatory privacy gateway that transforms personal thoughts into impersonal team knowledge.

Raw stays local

LLM distills

Team sync

Platform agnostic

Claude-Mem

Partial (<private> opt-out)

Cloud API compresses

Single-user

Claude Code only

Cipher

No

No

Yes

No

Supermemory

No

No

Yes

No

Mem0

Yes

No

No

Yes

Memctl

Yes

No

Yes

Yes

Distill

Yes

Yes

Yes

Yes

Based on public documentation as of March 2026.

Documentation

Development

git clone https://github.com/5queezer/distill.git
cd distill
uv sync
uv run pytest tests/ -x -v

License

MIT

Available Tools

8 tools
forgetA
Destructive

Soft-delete a memory. It will no longer appear in search results.

If agent_id is provided, the memory is only deleted when it belongs to that agent. Returns 'forbidden' if the memory belongs to a different agent.

Inputs: id (required), agent_id. Returns: object

get_lineageA
Read-only

Trace the supersedes chain for a memory in both directions.

Returns the full history: predecessors (what this memory replaced) and successors (what replaced this memory), ordered oldest to newest. Useful for understanding how a decision evolved over time.

Inputs: id (required). Returns: result

get_memoriesA
Read-only

Fetch full memory details by IDs. Batch multiple IDs in one call.

Use after search_memory to get content for relevant results only.

Inputs: ids (required). Returns: result

get_memoryB
Read-only

Retrieve a specific memory by its ID.

Inputs: id (required). Returns: object

list_recentA
Read-only

List recent memories as compact index. Use get_memories for full content.

Optionally filter by repo, tag, type, or agent_id.

Inputs: repo, tag, type, limit, +1 more. Returns: result

list_staleA
Read-only

List memories that are likely stale based on age and access patterns.

Stale memories have low Weibull survival scores and few accesses. Review the list and use forget() to clean up outdated knowledge.

Inputs: repo, limit. Returns: result

search_memoryA
Read-only

Search team knowledge before proposing architecture, creating files, refactoring, or answering "how should we..." questions.

Also search when the user says: "we decided", "last time", "previously", "remember when", "what's our pattern for".

Returns compact index (~30 tokens/result). Use get_memories to fetch full content for relevant IDs only (not all results). Optionally filter by repo name, agent_id, and/or date range (after/before as ISO 8601 strings, e.g. "2025-01-01").

Inputs: query (required), top_k, repo, agent_id, +2 more. Returns: result

update_memoryA

Re-distill new content and supersede an existing memory.

The old memory is soft-deleted. A new memory is created with the distilled version of the provided content.

Inputs: id (required), content (required). Returns: object

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A4.2/5.0

    Scored across 8 tools

    Disambiguation5/5

    Each tool targets a distinct aspect of memory management: search, retrieval by ID or batch, listing recent or stale, forgetting, updating with lineage, and tracing lineage. There is no ambiguity between tools.

    Naming Consistency5/5

    Tool names follow a consistent verb_noun pattern in snake_case (search_memory, get_memories, list_recent, update_memory, etc.), with 'list_' and 'get_' prefixes clearly indicating the action.

    Tool Count5/5

    Eight tools cover the core memory operations without being excessive or insufficient. Each tool serves a clear purpose in the memory lifecycle.

    Completeness4/5

    The surface covers most operations (search, retrieve, list, forget, update, lineage). The only noticeable gap is the lack of a dedicated create_memory tool, though update_memory might double as creation in some contexts.

    Maintenance

    ActivityNo data
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      A self-hosted MCP server that provides AI assistants with a shared, persistent SQLite-backed memory for storing and retrieving project context, decisions, and discoveries. It enables cross-session continuity and team-wide knowledge sharing to keep AI coding tools aligned and informed.
      3
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      MCP server that provides persistent memory and contextual awareness to language models, enabling project onboarding, recall of architectural rules, and code consistency across sessions.
      32
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      MCP server that captures and recalls coding session memory (failures, decisions, diffs) for AI agents, enabling cross-agent continuity and preventing repeated mistakes.
      69 npm
      MIT
    • A
      license
      Not graded
      quality
      C
      maintenance
      Self-hosted MCP server for storing and serving structured team knowledge, enabling AI sessions to load relevant team context without re-prompting.
      Apache 2.0