Zed
Add PromptsForLess to Zed as an OpenAI-compatible LLM provider and set each model's context window and capabilities.
Zed supports custom OpenAI-compatible providers in its LLM provider settings. The agent panel, inline assistant and edit predictions are configured separately, so a new provider does not change every AI feature.
Add PromptsForLess
- Open the command palette and run agent: open settings.
- In LLM Providers, choose Add Provider and select an OpenAI-compatible endpoint.
- Name it
PromptsForLessand enterhttps://api.promptsforless.com/v1as the API URL. - Add a model ID from the catalog, such as
deepseek-v4.1-flash. - Enter the model's context window and turn on only the capabilities it supports.
- Add your key from the dashboard in the provider settings, then pick the model in the agent panel.
Zed documents this flow in Use API Access. Add the key through Zed's UI rather than in settings.json.
Set capabilities to match the model
Leave the provider on Chat Completions. Zed's defaults for a compatible provider can assume tool support, so check each model's page in the catalog before an agent task. Turn on images, parallel tool calls, reasoning controls or prompt caching only for models that support them.
Test the connection
Send a short prompt in a new thread. Then try a read-only file task before allowing edits or terminal commands. To test the API outside Zed, send a direct request.
Troubleshooting
| Symptom | Fix |
|---|---|
| Provider missing | Check that the settings saved and the key was added. |
| Context errors | Set the context window to the value on the model's catalog page. |
| Tool or image requests fail | Turn off capabilities the model lacks, or pick another model. |
| Edit predictions use another service | Configure edit predictions separately in Zed. |
See the model catalog for prices and IDs.