Meerkat is an agent engine, not an agent product. Claude Code and Codex CLI are finished applications — Meerkat is what you’d build them on. It’s a set of modular Rust crates that handle the hard parts of agentic systems: state machines, retries, budgets, streaming, tool dispatch, MCP, multi-agent coordination. You bring your own prompts, tools, and opinions. Meerkat handles execution.Documentation Index
Fetch the complete documentation index at: https://docs.rkat.ai/llms.txt
Use this file to discover all available pages before exploring further.
Why Meerkat
- Fast. Rust from the ground up. Lean native binaries and embeddable libraries, <10ms cold start in the core path, and no VM/runtime dependency stack.
- Provider-agnostic. Same session/config model across Anthropic, OpenAI, and Gemini, with capability-gated differences where model support diverges.
- Modular. A workspace of focused crates, all opt-in. Need just the agent loop? Pull in
meerkat-core. Need session persistence? Addmeerkat-store. Need inter-agent comms? Addmeerkat-comms. Your binary includes only what you use. Disabled features return typed errors, not panics. - Stable. Deterministic state machine with defined transitions. Typed errors with compile-time guarantees. Every error code maps to a stable string across JSON-RPC, REST, and CLI exit codes.
- Controlled. Hard budget limits on tokens, wall-clock time, and tool calls. Eight hook points for guardrails and audit. Shell security with allow/deny lists.
- Multi-agent native. Independent agents communicate over Ed25519-signed channels, can be organized into mobs, and can keep working as compaction + semantic memory absorb conversations that exceed any single context window.
How it compares
| Meerkat | Claude Code / Codex CLI / Gemini CLI | |
|---|---|---|
| Design | Library-first, modular crates | CLI-first, SDK bolted on |
| Modularity | Compose only what you need | Monolithic, all-or-nothing |
| Performance | Single 5MB binary, <10ms startup, ~20MB RAM | Runtime + dependencies |
| Languages | Rust core + Python & TypeScript SDKs | TypeScript or Python |
| Surfaces | CLI, REST, JSON-RPC, MCP server, language SDKs | Rich interactive TUI |
| Multi-agent | Native Ed25519-authenticated P2P messaging | No |
| Memory | HNSW semantic indexing + auto-compaction | File-based context |
Use it from anywhere
Rust SDK
Embed as a library.
Python SDK
Drive from Python.
TypeScript SDK
Drive from TypeScript.
CLI
Run from the terminal.
REST API
HTTP endpoints.
JSON-RPC
Stateful stdio for IDEs.
Get started
The intended onboarding spine is:Introduction -> Quickstart -> Examples gallery -> choose a surface or workflow -> Concepts as needed
Quickstart
Get to a first successful run as fast as possible.
Examples gallery
The central real-code hub for choosing a language, surface, or workflow.
Auth setup
Configure credentials and bindings when you need more than the env-var fast path.
Choose your path
Rust embedding
Start here if Meerkat is part of your Rust application.
Python / TypeScript
Start here if you want an SDK-backed product layer.
CLI / APIs
Start here if you are operating Meerkat as a practical surface first.
Multi-agent systems
Jump here if your main interest is orchestration and long-lived identity-first agents.
Self-hosting models
Start here if your first concern is local or self-managed models.
Architecture
Start here if you need ownership, rationale, or internal structure first.
