Integration · ModelsUse DeepSeek models with FastClaw
Run FastClaw agents on DeepSeek — a built-in preset, thinking-mode support and per-agent model overrides for low-cost, capable agents.
One-line preset
The deepseek preset points at https://api.deepseek.com and reads DEEPSEEK_API_KEY from your environment.
Thinking mode handled
When DeepSeek requires reasoning_content to be passed back, FastClaw retries with it and sends it up front on later calls.
Cost-friendly defaults
Use DeepSeek as the default for busy chat bots and reserve pricier models for the agents that need them.
What you can do
- Power high-volume bots on Telegram or WeCom with a low-cost model.
- Mix DeepSeek with other model providers on a per-agent basis.
- Keep reasoning output for memory extraction while the agent works through tool calls.
Set it up
- Open Console → Models, click Add provider, choose DeepSeek and paste your API key.
- Pick a DeepSeek model as the default or set it on one agent's Models tab.
- From the CLI:
export DEEPSEEK_API_KEY=sk-...
fastclaw agents init helper \
--provider deepseek \
--model deepseek/deepseek-chat \
--api-key-env DEEPSEEK_API_KEY
Notes
Running a community bot on a budget? See FastClaw for communities. Want a fully local alternative? Try Ollama.
Frequently asked questions
More Integrations
All Integrations
OpenAI
Run FastClaw agents on OpenAI models — set a deployment-wide default, override per agent, and expose your agents through an OpenAI-compatible API.
Anthropic
Run FastClaw agents on Claude through Anthropic's native Messages API — with prompt caching, extended thinking and image input preserved.
Ollama
Point FastClaw at Ollama to run AI agents on local, open-weight models — no API key, no data leaving your machine.
OpenRouter
Connect FastClaw to OpenRouter and give every agent access to models from many vendors with a single API key.
Try it on your own agents
Start free in FastClaw Cloud, or self-host the single binary in a minute.