Now with Codex + MCP

Your agents change.
Your memory stays.

A self-hosted memory and knowledge layer for AI agents and humans. Carry project context across sessions, tools, and teams. Keep it in a wiki you can read, edit, and own.

  • Open source
  • Self-hosted
  • Your data
A SHARED KNOWLEDGE LAYER
NoosphereYour durable memory
  • CodexMCP + skill
  • OpenClawMemory plugin
  • HermesMemory provider
  • OpenCodeCoding agent
  • Kilo CodeCoding agent
  • YouA readable wiki
Different tools. The same project knowledge.

Start in OpenClaw.
Pick up in Codex. Keep the context.

One knowledge base
Scoped access
Readable by humans

01 / MEMORY THAT CARRIES FORWARD

Less starting over.
More building on what you know.

Decisions, discoveries, and the details that make a project yours. Give them a home that your next agent can find.

Recall the words.
Retrieve the meaning.

Search with PostgreSQL full-text retrieval out of the box. Add optional pgvector embeddings to find semantically related knowledge, even when the wording changes.

The memory layer ranks results, removes duplicates, handles conflicts, and fits useful context into your token budget.

Explore hybrid retrieval
ILLUSTRATIVE RECALL
Why did we choose PostgreSQL?
Keyword default Vector opt-in
Engineering / Decisions

Our storage architecture

We chose PostgreSQL to keep structured articles, full-text search, and optional vector retrieval together.

architecturedatabaseSource attached
Scoped to what this agent is allowed to read.
AUTO-CAPTURE & CURATION

Save the useful parts.
Keep control of the rest.

Choose how knowledge enters your memory. Agent saves create draft candidates by default, ready for review.

Explicit agent saves Available
Save a decision or discovery with a memory tool.
Plugin auto-save Optional
OpenCode and Kilo Code can save on idle when enabled.
Private capture API Opt-in
Scoped capture storage, lineage, and retention controls. Off by default.

Automatic turn extraction and automatic promotion are planned. View implementation status

A HUMAN HOME FOR AGENT KNOWLEDGE

Memory you can
actually work with.

Browse the same knowledge your agents use. Edit Markdown, organize nested topics, review revisions, and connect related articles.

WorkspaceEngineering↳ Decisions↳ RunbooksResearch
DECISION / REVIEWEDContext worth keeping. Revision history
Take it with you: Markdown + Obsidian

02 / UNDER THE SURFACE

One memory.
Connected, by design.

A shared article store, a provider-aware recall layer, and integrations that bring knowledge into the tools you already use.

  1. 01

    Your tools

    Agent plugins, Codex / MCP, REST clients, and the human wiki.

    CONNECT
  2. 02

    Memory layer

    Recall across providers. Rank, deduplicate, resolve conflicts, and budget context.

    ORCHESTRATE
  3. 03

    Hybrid retrieval

    Full-text search Default
    pgvector + RRF Opt-in
    FIND
  4. 04

    Durable knowledge

    PostgreSQL holds your articles and metadata. Redis supports fast recall caching.

    KEEP

Access stays scoped. Permissions apply to the knowledge each agent can retrieve.

Providers stay extensible. Recall supports Noosphere articles and Hindsight.

Vector retrieval combines lexical and semantic rankings with reciprocal rank fusion. It requires explicit activation, an embedding provider, and backfill. Keyword search remains the default.

Read the architecture

03 / BRING YOUR OWN AGENT

Your favorite tools.
A common memory.

Use a native plugin, connect through MCP, or build with the REST API. Every integration works with the same Noosphere instance.

Codex CLI + MCP

A little more context
for your next Codex session.

Connect Codex with a dedicated MCP server and memory skill. Search, recall, and save knowledge without leaving your workflow.

INSTALL THE CODEX INTEGRATION
npx -y @sweetsophia/noosphere-mcp@1.15.0 install-codex

Example pinned to v1.15.0. For a newer release, use the version recommended in the setup guide below.

Requires Node.js 22+, Codex CLI, a running Noosphere instance, and a scoped API key supplied through your environment.

search_articlesget_articlesave_memoryrecall_memorycreate_article
Set up Codex

Also works with other stdio MCP clients.

04 / MAKE IT YOURS

Give your next session
a head start.

Run Noosphere on your infrastructure.
Keep the knowledge you build along the way.

Read the installation guide Explore the source
  1. 01

    Start your instance

    The guided installer sets up the services. Start with Docker Compose v2 and Node.js 22+.

  2. 02

    Connect an agent

    Choose an integration and give it a scoped API key for the knowledge it needs.

  3. 03

    Build shared knowledge

    Save your first memory, review it in the wiki, and recall it in your next session.