Quickstart

Request Access

Self-hosted deployment is currently available by request. To get started, email us and we’ll provision a server for you with full access to the runtime, API, and dashboard.

Prerequisites

Before using Fathom, make sure you have:

• Server credentials — provided after your access request is approved
• API key — for the DeepSeek OpenAI-compatible endpoint (or another OpenAI-compatible endpoint)
• Optional: Chrome with –remote-debugging-port=9222 for browser CDP tools; Python/Node for REPL

First Run

Once you have server access, run your first autonomous worker task with a single command. Research is one built-in workflow; the same worker runtime also supports outreach, code, computer use and scheduled operations:

fathom run "Find B2B SaaS startups in Berlin with 10-50 employees"

Configure the LLM provider before running; Fathom does not prompt interactively for API keys. Results are saved to ./research-output/ by default.

Configuration Basics

Configuration is stored in ~/.fathom/config.toml. All sections are optional — missing fields use defaults.

Config Location

~/.fathom/config.toml

LLM Section

The only required configuration is your LLM provider settings:

[llm]
provider    = "deepseek"
base_url    = "https://api.deepseek.com"
api_key     = "sk-your-key"
model       = "deepseek-chat"
max_tokens  = 8192
temperature = 0.7

Search Backend

Choose a search engine for web queries. hybrid uses the first configured backend with results:

[search]
backend = "hybrid"

[search.tavily]
api_key = "tvly-your-key"

TUI Mode

Launch the interactive terminal UI for a real-time research experience with progress tracking, tool call visualization, and conversation history.

fathom tui

The TUI provides an interactive chat interface where you can type queries, watch agent progress in real time, and browse results.

HTTP Server Mode

Start an HTTP API server for programmatic access. Supports REST endpoints and SSE streaming for real-time event delivery.

fathom serve

The server starts on http://localhost:8080 by default. Send worker tasks via POST:

curl -X POST http://localhost:8080/api/v1/sessions \
-H "Content-Type: application/json" \
-d '{"query": "Find AI startups in London"}'

Next Steps

Now that you have Fathom running, explore these topics: