getcompanion-ai/co-mono
1
0
Fork 0
mirror of https://github.com/getcompanion-ai/co-mono.git synced 2026-04-17 05:00:16 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

v0.7.12: Custom models/providers support via models.json

Browse source 
- Add ~/.pi/agent/models.json config for custom providers (Ollama, vLLM, etc.)
- Support all 4 API types (openai-completions, openai-responses, anthropic-messages, google-generative-ai)
- Live reload models.json on /model selector open
- Smart model defaults per provider (claude-sonnet-4-5, gpt-5.1-codex, etc.)
- Graceful session fallback when saved model missing or no API key
- Validation errors show precise file/field info in CLI and TUI
- Agent knows its own README.md path for self-documentation
- Added gpt-5.1-codex (400k context, 128k output, reasoning)

Fixes #21
This commit is contained in:
Mario Zechner 2025-11-16 22:56:24 +01:00
parent 112ce6e5d1
commit 0c5cbd0068
15 changed files with 793 additions and 114 deletions
Download patch file Download diff file Expand all files Collapse all files

21
packages/coding-agent/CHANGELOG.md
View file

@ -2,6 +2,27 @@
## [Unreleased]
## [0.7.12] - 2025-11-16
### Added
- **Custom Models and Providers**: Support for custom models and providers via `~/.pi/agent/models.json` configuration file. Add local models (Ollama, vLLM, LM Studio) or any OpenAI-compatible, Anthropic-compatible, or Google-compatible API. File is reloaded on every `/model` selector open, allowing live updates without restart. ([#21](https://github.com/badlogic/pi-mono/issues/21))
- Added `gpt-5.1-codex` model to OpenAI provider (400k context, 128k max output, reasoning-capable).
### Changed
- **Breaking**: No longer hardcodes Anthropic/Claude as default provider/model. Now prefers sensible defaults per provider (e.g., `claude-sonnet-4-5` for Anthropic, `gpt-5.1-codex` for OpenAI), or requires explicit selection in interactive mode.
- Interactive mode now allows starting without a model, showing helpful error on message submission instead of failing at startup.
- Non-interactive mode (CLI messages, JSON, RPC) still fails early if no model or API key is available.
- Model selector now saves selected model as default in settings.json.
- `models.json` validation errors (syntax + schema) now surface with precise file/field info in both CLI and `/model` selector.
- Agent system prompt now includes absolute path to its own README.md for self-documentation.
### Fixed
- Fixed crash when restoring a session with a custom model that no longer exists or lost credentials. Now gracefully falls back to default model, logs the reason, and appends a warning message to the restored chat.
- Footer no longer crashes when no model is selected.
## [0.7.11] - 2025-11-16
### Changed