Connect a model
Interpreter does not ship with a model. You bring your own through a profile. A profile is one provider plus one model plus the credentials to reach it.
Profiles, not global settings
Models are configured per profile, not globally.
- Each profile stores its provider, model name, and key or base URL.
- You can keep as many profiles as you want — a hosted frontier model for hard work, a local model for sensitive files, a cheap model for batch jobs.
- Switching profiles mid-session is supported. The next turn uses the new profile.
See Profiles for the concept and how to set them up.
Supported providers
Out of the box:
- OpenAI — GPT-class models, direct.
- OpenRouter — multi-provider router, including Anthropic models.
- Ollama — local models on your machine.
- LM Studio — local models via the LM Studio server.
- NVIDIA NIM — hosted or self-hosted NIM endpoints.
- Any OpenAI-compatible endpoint — paste a base URL and key. This covers most self-hosted and third-party gateways.
For each one you need the model name and either an API key or, for local providers, a running server.
When to use which
Pick the smallest model that actually does the job. In practice:
- Hosted frontier models — use for judgment-heavy work: messy documents, ambiguous instructions, long multi-step browser flows, anything where the agent has to recover from surprises.
- Smaller hosted or local models — fine for repeatable structured extraction once you have a working instruction, simple file renaming, spreadsheet reformatting.
- Local models (Ollama, LM Studio) — best for sensitive workspaces, offline work, or cost-sensitive batch jobs. Expect slower turns and weaker recovery from unexpected UI state.
Validate before scaling
Local and smaller models often look correct on one example and drift on the tenth.
- Run on a single file first. Read the output.
- Run on five files. Spot-check.
- Then run on the full batch.
If a model keeps making the same mistake, switch profiles to a stronger model for that run rather than fighting the prompt. You can switch back after.
Next: Profiles.