feat: add runtime provider/model switching via /model <tier> <provider/model>

- ModelRouter: add setClient(), labels map, getLabel(), getAllLabels()
- TUI commands: parse /model <tier> <provider/model> syntax with autocompletion
- TUI minimal: handle provider switching via createClientFromConfig factory
- Daemon: wire initial labels into router config
- Fix /model alias mappings (opus=complex, sonnet=default, haiku=fast)
- Add design doc and update state.json with feature status
This commit is contained in:
William Valentin
2026-02-06 23:42:14 -08:00
parent e92ce69067
commit d4530a7034
8 changed files with 527 additions and 37 deletions
@@ -0,0 +1,79 @@
# Provider/Model Runtime Switching
**Date:** 2026-02-06
**Status:** In Progress
## Goal
Enable easy runtime switching of model providers per tier via the `/model` command, using `provider/model` syntax.
## Commands
```
/model — Show all tiers with provider/model labels
/model fast — Switch active tier (existing behavior)
/model default github-copilot/claude-sonnet-4-5 — Change default tier's provider+model
/model complex anthropic/claude-opus-4 — Change complex tier's provider+model
/model fast github-copilot/gpt-4o-mini — Change fast tier's provider+model
```
## Design Decisions
1. **No presets** — Direct `provider/model` targeting per tier. YAGNI.
2. **Full override** — When you set a tier, it fully replaces the previous client.
3. **Local tier excluded**`/model local` continues to use `/backend` for switching. Local models are a different concern.
4. **Auth** — Config-based `api_key` for most providers. `/login` OAuth flow for GitHub Copilot (already implemented).
5. **Merged into /model** — No separate `/provider` command. Everything lives under `/model`.
## Implementation
### 1. ModelRouter (`src/models/router.ts`)
- Add `setClient(tier: ModelTier, client: ModelClient, label: string)` — replaces a tier's client at runtime
- Add `getLabel(tier: ModelTier): string` — returns `provider/model` string for display
- Track labels in a `Map<ModelTier, string>` populated at construction and updated by `setClient()`
### 2. Client Factory (`src/daemon/index.ts`)
Extract `createClientFromProvider(provider: string, model: string, opts?: { apiKey?: string; endpoint?: string }): ModelClient` factory function from the existing inline client creation logic. Used by both daemon startup and runtime `/model` switching.
### 3. Command Parser (`src/frontends/tui/commands.ts`)
- Extend `Command` type: `{ type: 'model'; name?: string; providerModel?: string }`
- Parse `/model <tier> <provider/model>` — split on space to get tier + provider/model
- Parse provider/model string: split on first `/` to get provider and model name
- Update autocompletion to suggest available providers after tier name
- Update tooltips
### 4. Daemon Wiring (`src/daemon/index.ts`)
Handle the new command variant:
1. Receive `{ type: 'model', name: 'default', providerModel: 'github-copilot/claude-sonnet-4-5' }`
2. Parse provider and model from `providerModel`
3. Call `createClientFromProvider(provider, model)` to instantiate client
4. Call `router.setClient('default', client, 'github-copilot/claude-sonnet-4-5')`
5. Respond with confirmation message
### 5. Provider Name Mapping
Map short provider names to client constructors:
| Provider Name | Client Class |
|---------------|-------------|
| `anthropic` | AnthropicClient |
| `openai` | OpenAIClient |
| `github` / `github-copilot` | GitHubModelsClient |
| `gemini` | GeminiClient |
| `bedrock` | BedrockClient |
| `ollama` | OllamaClient |
| `llamacpp` | LlamaCppClient |
## Files Changed
| File | Change |
|------|--------|
| `src/models/router.ts` | Add `setClient()`, `getLabel()`, label tracking |
| `src/models/router.test.ts` | Tests for new methods |
| `src/frontends/tui/commands.ts` | Extended parser, completions, tooltips |
| `src/frontends/tui/commands.test.ts` | Tests for new parsing |
| `src/daemon/index.ts` | Extract factory, wire new command handler |
+135 -2
View File
@@ -336,6 +336,136 @@
}
}
},
"p4-media-pipeline": {
"status": "completed",
"date": "2026-02-06",
"summary": "Multimodal media pipeline: receive images from channel adapters and pass through to vision-capable models (Anthropic, OpenAI, Gemini, Bedrock)",
"phases": {
"type_widening": {
"priority": "P4",
"status": "completed",
"description": "Widen Message.content from string to string | MessageContentPart[], add Attachment type to channel layer, add ImageSource/MessageContentPart types",
"files_created": [
"src/models/media.ts",
"src/models/media.test.ts"
],
"files_modified": [
"src/models/types.ts",
"src/models/index.ts",
"src/channels/types.ts",
"src/channels/index.ts"
],
"test_status": "25/25 passing"
},
"model_client_multimodal": {
"priority": "P4",
"status": "completed",
"description": "Update all model clients to convert MessageContentPart[] to provider-specific image formats (Anthropic base64, OpenAI data URI, Gemini inlineData, Bedrock image bytes)",
"files_modified": [
"src/models/anthropic.ts",
"src/models/openai.ts",
"src/models/gemini.ts",
"src/models/bedrock.ts",
"src/models/local/llamacpp.ts",
"src/models/local/ollama.ts"
]
},
"agent_attachment_passthrough": {
"priority": "P4",
"status": "completed",
"description": "Wire attachments through NativeAgent.process() and AgentOrchestrator.process() to daemon message handler",
"files_modified": [
"src/backends/native/agent.ts",
"src/backends/native/orchestrator.ts",
"src/daemon/index.ts"
]
},
"downstream_type_fixes": {
"priority": "P4",
"status": "completed",
"description": "Fix all consumers of Message.content to use getMessageText() helper: token estimation, compaction, TUI rendering",
"files_modified": [
"src/context/tokens.ts",
"src/context/compaction.ts",
"src/frontends/tui/components/MessageList.tsx"
]
},
"channel_adapter_extraction": {
"priority": "P4",
"status": "completed",
"description": "Extract images from platform messages in all channel adapters",
"sub_phases": {
"telegram": {
"status": "completed",
"description": "Handle message:photo (largest size, download via getFile API, base64) and image message:document events with caption text",
"files_modified": ["src/channels/telegram/adapter.ts"]
},
"discord": {
"status": "completed",
"description": "Extract image attachments from message.attachments Collection, pass Discord CDN URLs directly",
"files_modified": ["src/channels/discord/adapter.ts"]
},
"slack": {
"status": "completed",
"description": "Download image files via url_private_download with bot token auth, base64 encode",
"files_modified": ["src/channels/slack/adapter.ts"]
},
"whatsapp": {
"status": "completed",
"description": "Use downloadMedia() from whatsapp-web.js (returns base64 natively)",
"files_modified": ["src/channels/whatsapp/adapter.ts"]
},
"webchat": {
"status": "deferred",
"description": "Requires gateway protocol update for WebSocket attachment messages"
}
}
}
}
},
"p5-github-copilot-provider": {
"status": "completed",
"date": "2026-02-06",
"summary": "GitHub Copilot as a model provider with OAuth device flow and auto-login on first use",
"phases": {
"copilot_client": {
"priority": "P5",
"status": "completed",
"description": "GitHubModelsClient using OpenAI SDK against api.githubcopilot.com with Copilot-specific headers, multimodal support, streaming, tool calls",
"files_created": [
"src/models/github.ts",
"src/auth/github.ts",
"src/auth/index.ts"
],
"files_modified": [
"src/config/schema.ts",
"src/models/index.ts",
"src/models/costs.ts",
"src/daemon/index.ts",
"src/cli/tui.ts"
]
},
"oauth_device_flow": {
"priority": "P5",
"status": "completed",
"description": "Interactive OAuth device flow via /login github command, token stored at ~/.config/flynn/auth.json with chmod 0600",
"files_modified": [
"src/frontends/tui/minimal.ts",
"src/frontends/tui/commands.ts"
]
},
"auto_login": {
"priority": "P5",
"status": "completed",
"description": "Lazy token resolution with onLoginRequired callback — triggers OAuth device flow automatically on first API call when no token is available",
"files_modified": [
"src/models/github.ts",
"src/daemon/index.ts",
"src/cli/tui.ts"
]
}
}
},
"earlier_plans": {
"status": "completed",
"summary": "Original design and implementation phases from 2026-02-02 to 2026-02-05",
@@ -361,11 +491,14 @@
},
"overall_progress": {
"total_test_count": 655,
"total_test_count": 742,
"all_tests_passing": true,
"p0_completion": "3/3 (100%)",
"p1_completion": "4/4 (100%)",
"p2_completion": "7/7 (100%)",
"next_up": "p3_remaining (group chat support, gateway auth, gemini provider, browser control, additional model providers)"
"p3_completion": "completed (group chat, gateway auth, Gemini, OpenRouter, Bedrock, browser control)",
"p4_completion": "1/1 (100%) — multimodal media pipeline",
"p5_completion": "1/1 (100%) — GitHub Copilot provider with auto-login",
"next_up": "p6 (image.analyze tool, audio transcription, outbound attachments, gateway protocol attachments)"
}
}