OpenClaw on SaladCloud
Your OpenClaw agent. 86% cheaper.
Stop paying per token. Run OpenClaw against an OpenAI-compatible endpoint on SaladCloud's flat-rate GPUs from $0.09 per hour, or through AI Gateway per token. Add workflows, tasks, and agents without watching the meter.
No contracts. No minimums. Pay only while your endpoint is running.
# openclaw config — point the model provider at your Salad endpoint providers: salad: type: openai base_url: https://<your-container-gateway>.salad.cloud/v1 api_key: ${SALAD_ENDPOINT_KEY} model: qwen3.6-35b-a3b default_provider: salad
Three steps
From per-token billing to flat-rate compute.
Pick a recipe and deploy
Choose Ollama, vLLM, or TGI in the portal, select a GPU, and get an OpenAI-compatible endpoint in minutes.
Point OpenClaw at Salad
Add the endpoint as a model provider in your OpenClaw config and set it as the default for automated work.
Watch the bill flatten
Your agent runs on a fixed hourly GPU. More workflows, more tasks, more agents; same invoice.
Prefer not to run a model at all? AI Gateway serves Qwen-class models per token or on a flat monthly subscription, and works with OpenClaw the same way.
Resources