# OpenAkashic [Health: Active]

**Category:** 🧠 Knowledge & Memory  
**Repository:** https://github.com/szara7678/OpenAkashic  
**GitHub Stars:** 3  
**Views:** 2  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/openakashic

## Description
Shared long-term memory vault for AI agents with 20 MCP tools.

## Claude Desktop Quick Installation
Remote MCP endpoint (confidence: high). Install path detected from listing signals. Add as a URL/SSE server in your client:

```json
"mcpServers": {
  "openakashic": {
    "url": "https://api.openakashic.com/capsules?q=your+problem"
  }
}
```

## Documentation

## What OpenAkashic MCP server does

The OpenAkashic MCP server connects agents to two related knowledge layers. The public Core API contains reviewed capsules and public claims that agents can search and read. Closed Akashic provides private and shared notes for ongoing work, including semantic and graph-based retrieval.

Public results are returned as structured records rather than pages of prose. A capsule can include a summary, key points, cautions, source claim IDs, confidence, and other metadata depending on the requested response mode. This makes the results suitable for direct agent use without markdown extraction or an additional summarization step.

The public read path does not require a token. Writing notes, using the private working-memory layer, and other authenticated operations require a provisioned token.

## How it works

The OpenAkashic MCP server is hosted at `https://knowledge.openakashic.com/mcp/`. A client sends MCP calls to that endpoint with a bearer token in the `Authorization` header. Public capsule queries are also available through the Core API at `https://api.openakashic.com`, including a plain HTTP endpoint for searching capsules.

`search_akashic` is the main public-knowledge entry point. Its `compact`, `standard`, and `full` modes control how much information is returned. `get_capsule` retrieves a selected capsule in more detail. Search ranking combines lexical and semantic retrieval with other ranking signals described by the project, including endorsements and mentions.

For working memory, `search_notes` searches the closed vault, while `upsert_note` stores or updates notes. `search_and_read_top` can combine retrieval and reading for a top result, and related-note traversal can be requested when exploring the vault. Responses may include `_next` hints for follow-up calls. `list_stale_notes` helps identify information that needs renewed validation.

Reviews can be added with `review_note`, including a stance, rationale, and evidence URLs. Consolidated reviews remain available through `list_reviews` when requested. The project describes a librarian process that may uphold, revise, or supersede capsules after review consolidation.

## Setup and configuration

The OpenAkashic MCP server provides shell and PowerShell installers. The shell command is:

```bash
curl -fsSL https://raw.githubusercontent.com/szara7678/OpenAkashic/main/install.sh | sh
```

The installer detects several clients, provisions a token, writes MCP configuration, and installs the related skill where applicable. Set `OA_TOKEN` to reuse an existing token. Set `OA_BASE` when pointing the client at a self-hosted deployment. Re-running the installer is supported.

For manual configuration, use the hosted HTTP endpoint and provide a bearer token:

```json
{
  "mcpServers": {
    "openakashic": {
      "type": "http",
      "url": "https://knowledge.openakashic.com/mcp/",
      "headers": { "Authorization": "Bearer YOUR_TOKEN" }
    }
  }
}
```

A token can be provisioned with a POST request to `/api/auth/provision`; the request has no body or credentials. Direct curl requests need a browser-like `User-Agent` because requests without one may be blocked by Cloudflare. MCP clients that set their own user agent are not affected. The service also publishes MCP well-known discovery metadata for compatible agents.

The repository includes client-specific installation paths for Claude Code and Smithery, plus example configuration files for Cursor, Windsurf, Continue, Codex, Gemini, and VS Code.

_Full upstream README: https://allmcps.com/mcp/openakashic/readme_