Bodega One beta.45 is out. - Desktop control on Windows: asks first, highlights every action, Esc to stop - Smarter agent: fewer wasted steps, cheaper long runs, steer it mid-reply - New models: Sonnet/Haiku 5.5, GPT-6.1 Sol, Gemini 3.6 Flash - Calmer UI, smoother streaming, one-click background updates - Instruction sets, per-chat memory, safer by default

1/ Bodega One beta.42 is out. Pick Claude Code, Codex, Gemini CLI or Cursor from the composer and the turn runs there, with Bodega's memory, rules and approvals around it, and a review card at the end showing what changed and whether your tests still pass.

Stay tuned for the next update, soon you can bring in any Harness into Bodega one Code, Setup your main & sub-agents with their own Harness/Provider/Model, and agents will always retain full context.

Your local small model is struggling? Soon in Bodega one code your models will be able to spawn sub agents with their own custom persona and be able to spawn models such as Fable 5.1 (Via Byok) or other bigger local models to finish the work it can’t.

Bodega One v1.0.0-beta.39 is out. This one's about honesty. When the agent gets cut off mid-task, it now says so, in the app and the CLI — instead of presenting half an answer as a finished one. Long local sessions stop losing your original request when memory gets tight. And routing now does exactly what the settings page says it does. Also: Model Roles is four pickers instead of eight, attachments stopped tripping false "too long" errors, and we fixed a filter that was silently eating angle brackets from your messages, pasted code keeps its generics now. Free . bodegaone.ai

It asked before it opened a browser. Everything on this screen is Bodega One running locally — a 27B model writing code, showing the diff, waiting for permission. No cloud, no API key, no data leaving the desk.

Cooking up something awesome, So Bodega Code already supposed calling in vision models as a tool for text only models, now we are introducing Bodega Agent Browser, the agent can open any webpage, search, click, basically use it as a normal human and take screenshots if needed, This was tested with Qwen3.6: 27B and text only model and it was able to call in a small vision model read the contents of the page and take screenshots. @BodegaOneAI Is turning out to be out the best Local First IDE, 100% Air-gap and private with optional Cloud via BYOK Stay tuned for the new upcoming update.