About Aurora Grimoire
Aurora Grimoire is an MCP server in the RAG category: mCP server for Aurora OS documentation retrieval. Feeds your RAG pipeline with relevant, up-to-date snippets from https://developer.auroraos.ru/ — no fluff, no hallucinations.
Install
git clone https://github.com/KotDath/aurora-grimoireThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Ask AI
Ask AI about Aurora Grimoire
Powered by Claude · Grounded in docs
Security checks
- ✓README presentThe repository ships documentation.
5 checks have not been run yet.
README
Aurora Grimoire
CLI + MCP server for Aurora OS documentation RAG.
The project covers the full pipeline:
- fetch HTML documentation;
- normalize into Markdown;
- chunk content;
- generate embeddings;
- upload vectors to Qdrant;
- search with sources (
cli search_docs); - expose search through MCP tool
search_docs.
Features
- Full offline-friendly RAG pipeline via CLI.
- Hybrid retrieval: dense (Qdrant) + BM25.
- Documentation version filtering (
--doc-version). - Knowledge sufficiency threshold (
--knowledge-threshold). - Artifact transfer between machines via bundle.
- MCP server (
stdioand HTTP transport).
Requirements
- Rust/Cargo (to install CLI via
cargo install --path .). - For
rag embed: Ollama with an embedding model. - For
rag deployand search: Qdrant.
For a local dev stack:
rag dev upstarts Docker services (Qdrant + Ollama).
Installation
Install the CLI binary:
cargo install --path .
Check installation:
aurora-grimoire --version
If command is not found, ensure Cargo bin is in PATH:
export PATH="$HOME/.cargo/bin:$PATH"
Config Is Required
All commands except rag config init require an existing config.toml.
Create config:
aurora-grimoire rag config init
Overwrite existing config:
aurora-grimoire rag config init --force
Custom config path:
AURORA_GRIMOIRE_CONFIG=/path/to/config.toml aurora-grimoire rag config init
Config template: config.example.toml
Quick Start (end-to-end)
- Initialize config:
aurora-grimoire rag config init
- Start local stack:
aurora-grimoire rag dev up --build --gpu --verbose
- Fetch and prepare data:
aurora-grimoire rag fetch-web -v
aurora-grimoire rag struct -v
aurora-grimoire rag chunk -v
- Generate embeddings and deploy to Qdrant:
aurora-grimoire rag embed -v
aurora-grimoire rag deploy -v --recreate
- Validate search:
aurora-grimoire cli search_docs "how to build project with mb2" --doc-version 5.2.0 --json
Transfer Artifacts To Another User
rag bundle create always creates a dual bundle (chunks + vectors).
Create bundle:
aurora-grimoire rag bundle create --out ~/.aurora-grimoire/bundles/aurora-dual.tar.zst
Inspect bundle:
aurora-grimoire rag bundle inspect --file ~/.aurora-grimoire/bundles/aurora-dual.tar.zst
On another machine:
aurora-grimoire rag config init
aurora-grimoire rag bundle extract --file /path/aurora-dual.tar.zst
aurora-grimoire rag deploy --input ~/.aurora-grimoire/vectors_data --recreate -v
MCP
Start MCP server:
aurora-grimoire mcp start --stdio
or HTTP:
aurora-grimoire mcp start --http --host 127.0.0.1 --port 8080
MCP smoke test:
aurora-grimoire mcp smoke --doc-version 5.2.0 --query "how to build project with mb2"
MCP exposes one tool:
search_docs
Agent Integrations
Ensure aurora-grimoire is installed and available in PATH.
If you need absolute binary path:
AURORA_GRIMOIRE_BIN="$(command -v aurora-grimoire)"
Use that absolute path in examples below.
Claude Code
Stdio transport:
claude mcp add aurora-grimoire -- $AURORA_GRIMOIRE_BIN mcp start --stdio
HTTP transport:
aurora-grimoire mcp start --http --host 127.0.0.1 --port 8080
claude mcp add --transport http aurora-grimoire-http http://127.0.0.1:8080/mcp
Codex
Stdio transport:
codex mcp add aurora-grimoire -- $AURORA_GRIMOIRE_BIN mcp start --stdio
HTTP transport:
aurora-grimoire mcp start --http --host 127.0.0.1 --port 8080
codex mcp add aurora-grimoire-http --url http://127.0.0.1:8080/mcp
Check:
codex mcp list
OpenCode
OpenCode adds MCP server via interactive wizard:
opencode mcp add
For local mode set:
name:aurora-grimoiretransport/type:local/stdiocommand:$AURORA_GRIMOIRE_BINargs:mcp start --stdio
For remote mode set:
transport/type:remote/httpurl:http://127.0.0.1:8080/mcp
Check:
opencode mcp list
Other MCP Clients
If your client supports stdio mcpServers config, use:
{
"mcpServers": {
"aurora-grimoire": {
"command": "$AURORA_GRIMOIRE_BIN",
"args": ["mcp", "start", "--stdio"]
}
}
}
If your client supports streamable HTTP, start server:
aurora-grimoire mcp start --http --host 127.0.0.1 --port 8080
and connect URL http://127.0.0.1:8080/mcp.
Install Agent Skill/Command Templates
CLI includes installer command:
aurora-grimoire agents install --runtime all --scope global --verbose
Flags:
--runtime <claude|opencode|codex|all>(default:all)--scope <global|local>(default:local)--config-dir <PATH>(single runtime only)--force(overwrite existing files)
Installed targets:
- Claude Code:
commands/aurora/search-docs.md(invoke:/aurora:search-docs ...) - OpenCode:
command/aurora-search-docs.md(invoke:/aurora-search-docs ...) - Codex:
skills/aurora-search-docs/SKILL.md(invoke:$aurora-search-docs ...)
Main Commands
aurora-grimoire
├─ agents
│ └─ install
├─ rag
│ ├─ config init
│ ├─ fetch-web
│ ├─ struct
│ ├─ chunk
│ ├─ embed
│ ├─ deploy
│ ├─ bundle (create/inspect/extract)
│ ├─ dev (up/down/status/logs)
│ ├─ test-e2e
│ └─ clear
├─ cli
│ └─ search_docs
└─ mcp
├─ start
└─ smoke
Full and current command tree with flags: docs/architecture.md
Useful Commands
Check dev stack status:
aurora-grimoire rag dev status
Logs:
aurora-grimoire rag dev logs -f
Clear artifacts:
aurora-grimoire rag clear --all
README copy captured when this server was indexed; it may lag the repository. The original is authoritative.