LM Studio
LM Studio is a first-class Agent > Models provider for local open-weight models. Fased talks to LM Studio through its OpenAI-compatible chat endpoint and discovers models from the LM Studio local model API. Fased does not install LM Studio and does not download its model weights. Run LM Studio separately, then connect its local API from Agent > Models. Default endpoints:- Chat API:
http://127.0.0.1:1234/v1 - Model discovery:
http://127.0.0.1:1234/api/v1/models
Quick start
- Install LM Studio and start the local server.
- Load or download a model in LM Studio.
- Open Agents > selected Agent > Models > LM Studio.
- Enter
http://127.0.0.1:1234/v1, the model key, and optional API token. - Assign the model to the Agent’s primary/fallback/task roles.
Finding the model key
Ask LM Studio for its model list:key or id as the model ID. Fased model refs look like:
Config shape
Security
LM Studio is treated as a private-network provider by default. Keep it on localhost or a trusted tailnet/LAN host. If you enable LM Studio API authentication, save the token in Agent > Models or expose it throughLM_API_TOKEN.