fleet-rlm 0.4.95

Recursive Language Models with DSPy + Modal and an integrated Web UI for secure long-context code execution

fleet-rlm

Secure, cloud-sandboxed Recursive Language Models (RLM) with DSPy and Modal.

fleet-rlm gives AI agents a secure cloud sandbox for long-context code and document work, with a Web UI-first experience, recursive delegation, and DSPy-aligned tooling.

Paper | Docs | Contributing

---

Quick Start

Install and launch the Web UI in under a minute:

# Option 1: install as a runnable tool
uv tool install fleet-rlm
fleet web

Or in your active environment:

# Option 2: regular environment install
uv pip install fleet-rlm
fleet web

Open http://localhost:8000 in your browser.

fleet web is the primary interactive interface. The published package already includes the built frontend assets, so end users do not need bun or a separate frontend toolchain.

What You Get

• Browser-first RLM chat (fleet web)
• Secure Modal-backed long-context execution for code/doc workflows
• WS-first runtime streaming for chat and execution events
• Runtime configuration and diagnostics from the Web UI settings
• Optional MCP server surface (fleet-rlm serve-mcp)

Common Commands

# Standalone terminal chat
fleet-rlm chat --trace-mode compact

# Explicit API server
fleet-rlm serve-api --port 8000

# MCP server
fleet-rlm serve-mcp --transport stdio

# Scaffold assets for Claude Code
fleet-rlm init --list

Runtime Notes

• Product chat transport is WS-first (/api/v1/ws/chat).
• Runtime model updates from Settings are hot-applied in-process (/api/v1/runtime/settings) and reflected on /api/v1/runtime/status.
• Secret inputs in Runtime Settings are write-only.

Running From Source (Contributors)

# from repo root
uv sync --extra dev --extra server
uv run fleet web
uv run fastapi dev

For release/packaging workflows, uv build now runs frontend build sync automatically (requires bun in repo checkouts that include src/frontend).

Use full contributor setup and quality gates in AGENTS.md and CONTRIBUTING.md.

Architecture Overview

Read this after the quick start if you want the full system picture (entry points, ReAct orchestration, tools, Modal execution, persistent storage).

graph TB
    subgraph entry ["🚪 Entry Points"]
        CLI["fleet / fleet-rlm CLI"]
        WebUI["Web UI<br/>(React SPA)"]
        API["FastAPI<br/>(WS/REST)"]
        TUI["Ink TUI<br/>(standalone runtime)"]
        MCP["MCP Server"]
    end

    subgraph orchestration ["🧠 Orchestration Layer"]
        Agent["RLMReActChatAgent<br/>(dspy.Module)"]
        LMs["Planner / Delegate LMs"]
        History["Chat History"]
        Memory["Core Memory<br/>(Persona/Human/Scratchpad)"]
        DocCache["Document Cache"]
    end

    subgraph tools ["🔧 ReAct Tools"]
        DocTools["📄 load_document<br/>read_file_slice<br/>chunk_by_*"]
        RecursiveTools["🔄 rlm_query<br/>llm_query<br/>(recursive delegation)"]
        ExecTools["⚡ execute_code<br/>edit_file<br/>search_code"]
    end

    subgraph execution ["⚙️ Execution Layer"]
        Interpreter["ModalInterpreter<br/>(JSON protocol)"]
        Profiles["Execution Profiles:<br/>ROOT | DELEGATE | MAINTENANCE"]
    end

    subgraph cloud ["☁️ Cloud & Persistence"]
        Sandbox["Modal Sandbox<br/>(Python REPL + Driver)"]
        Volume[("💾 Modal Volume<br/>/data/<br/>• workspaces<br/>• docs/metadata")]
        Neon[("🐘 Neon Postgres<br/>• runs / steps<br/>• artifacts<br/>• tenants")]
        PostHog["📈 PostHog<br/>(LLM Observability)"]
    end

    WebUI -->|"WS / REST"| API
    CLI --> Agent
    API --> Agent
    TUI --> Agent
    MCP --> Agent

    Agent --> LMs
    Agent --> History
    Agent --> Memory
    Agent --> DocCache

    Agent --> DocTools
    Agent --> RecursiveTools
    Agent --> ExecTools

    API -.->|"Persistence"| Neon
    Agent -.->|"Traces"| PostHog

    DocTools --> Interpreter
    RecursiveTools --> Interpreter
    ExecTools --> Interpreter

    Interpreter --> Profiles
    Interpreter -->|"stdin/stdout<br/>JSON commands"| Sandbox
    Sandbox -->|"read/write"| Volume

    style entry fill:#e3f2fd,stroke:#1976d2,stroke-width:2px
    style orchestration fill:#f3e5f5,stroke:#7b1fa2,stroke-width:2px
    style tools fill:#fff3e0,stroke:#f57c00,stroke-width:2px
    style execution fill:#e8f5e9,stroke:#388e3c,stroke-width:2px
    style cloud fill:#fce4ec,stroke:#c2185b,stroke-width:2px

Docs and Guides

• Documentation index
• Explanation index
• Quick install + setup
• Configure Modal
• Runtime settings (LM/Modal diagnostics)
• Deploying the server
• Using the MCP server
• CLI reference
• HTTP API reference
• Auth modes
• Database architecture
• Source layout

Advanced Features (Docs-First)

fleet-rlm also supports runtime diagnostics endpoints, WebSocket execution streams (/api/v1/ws/execution), multi-tenant Neon-backed persistence, and opt-in PostHog LLM analytics. Those workflows are documented in the guides/reference docs rather than front-loaded here.

Contributing

Contributions are welcome. Start with CONTRIBUTING.md, then use AGENTS.md for repo-specific commands and quality gates.

License

MIT License — see LICENSE.

Based on Recursive Language Modeling research by Alex L. Zhang (MIT CSAIL), Omar Khattab (Stanford), and Tim Kraska (MIT).