Agent Booster
by sseshachalacommunityv0.6.4
1
Agent Booster is an MCP server and CLI that indexes code symbols in a project, builds optional local embeddings, and serves tools that return only relevant functions/classes for a task instead of whole files. It also provides setup/teardown commands for MCP clients like Claude Code, Cursor, Windsurf, and Codex, plus token-savings reporting.
From the registry: Cut AI agent token costs 5-15x — routes only relevant code symbols instead of full files.
Install & configure
Prerequisites
- requiredInstall Python 3.10+Python 3.10 or newer is required.
- requiredInstall Agent BoosterInstall the MCP server package. The embed extra is recommended because semantic search features use embeddings.
$ pip install agent-booster[embed] - requiredIndex the project codebaseRun once from the project root to build the symbol index used by the server.
$ booster index - requiredBuild embeddingsRecommended and effectively required for the default semantic search experience described in the README.
$ booster embed
Install in your client
Paste into the AI you're chatting with — it will figure out the rest
Please install the `agent-booster` MCP server into my current AI client (that's you).
Required prerequisites (do these first if not already done):
- **Install Python 3.10+** — Python 3.10 or newer is required.
- **Install Agent Booster** — Install the MCP server package. The embed extra is recommended because semantic search features use embeddings. Run: `pip install agent-booster[embed]`
- **Index the project codebase** — Run once from the project root to build the symbol index used by the server. Run: `booster index`
- **Build embeddings** — Recommended and effectively required for the default semantic search experience described in the README. Run: `booster embed`
Canonical MCP server config (stdio transport):
- command: `booster`
- args: ["serve"]
Note: No environment variables are required. The server runs over stdio via `booster serve`. The README also shows platform-specific `booster init <platform>` helpers, but the canonical MCP config shape is simply command `booster` with args `["serve"]`.
Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.Note: No environment variables are required. The server runs over stdio via `booster serve`. The README also shows platform-specific `booster init <platform>` helpers, but the canonical MCP config shape is simply command `booster` with args `["serve"]`.
Tools (4)
get_symbols
Returns all indexed symbols for a file, including name, kind, line range, and signature.
search_context
Performs semantic search across indexed symbols to find relevant code for a task, falling back to keyword search if embeddings are unavailable.
smart_read
Returns only the source lines for symbols in a file that match the given task, rather than the entire file.
route_model
Recommends a model choice (haiku, sonnet, or opus) based on task complexity and optionally the set of relevant files.
Prerequisites
- • Python 3.10+
Packages
agent-booster
pypiv0.6.4transport: stdio
Details
- Language
- Python
- License
- MIT
- Released
- Jun 27, 2026
- Last commit
- Jun 27, 2026
- Registry name
- io.github.sseshachala/agent-booster
Categories
Related Servers
Claude Flow
AI orchestration with hive-mind swarms, neural networks, and 87 MCP tools for enterprise dev.
65.5k
Codebase Memory Mcp
Persistent codebase knowledge graph. Survives session restarts and context compaction.
33.8k
Serena
A powerful toolkit for coding, providing semantic retrieval and editing capabilities.
26.7k