mirror of
https://github.com/bytedance/deer-flow.git
synced 2026-08-10 14:58:46 +00:00
* feat(subagents): persist and display subagent step history (#3779) Capture both assistant turns and tool outputs during subagent execution, stream them in task_running events, and persist them as subagent.* run events so the subtask card's step timeline survives a reload. Backend: - step_events.py: pure layer (capture_step_message, build_subagent_step, subagent_run_event) shared by streaming and persistence - executor.py: capture ToolMessage outputs, not just AIMessage turns - worker.py: persist task_* custom events to RunEventStore (category "subagent" keeps them out of the thread feed; list_events backfills) Frontend: - core/tasks/steps.ts + api.ts: SubtaskStep model, messageToStep, eventsToSteps, mergeSteps, fetchSubtaskSteps - subtask card accumulates live steps and backfills on expand - carry run_id onto history content messages for the events endpoint * fix(subagents): show AI turns in subtask card + paginate step backfill (#3779) Two follow-ups to the subagent step-history feature: Problem 1 — reload backfill could silently truncate the step timeline because list_events capped at 500 events (seq-ASC) across the whole run. Add task_id filtering + an after_seq forward cursor to list_events (all three stores + abstract base + the /events route), and make fetchSubtaskSteps page through one task's subagent.step events until a short page. No schema migration: the DB filter rides the existing run-scoped index via event_metadata["task_id"]. Problem 2 — the card only rendered tool steps, so persisted AI turns were never shown. Replace toolStepsForDisplay with stepsForDisplay: interleave AI reasoning turns (with text) and tool steps by message_index, drop blank-text AI turns, and drop the trailing final-answer AI turn when completed (already shown as result). Card renders AI steps as muted clamped markdown with a sparkles icon. Tests: store task_id/after_seq filtering + pagination across memory/db/jsonl, the /events route forwarding, stepsForDisplay rules, and fetchSubtaskSteps pagination. Docs updated in both AGENTS.md. * make format * fix(subagents): capture full multi-tool step tail, batch step persistence, cap tool-call args (#3779) Address PR review findings on the subagent step-history feature: 1. executor.py streamed on stream_mode="values" and captured only messages[-1] per chunk, so a multi-tool-call turn (ToolNode appends one ToolMessage per call in a single super-step) lost all but the last tool output in both the live task_running stream and the persisted history. Replace with capture_new_step_messages, which walks the newly-appended tail (and still re-checks the trailing message on no-growth chunks so id-less in-place replacements survive). 2. worker.py persisted each step with the store's low-frequency put() (a per-thread advisory lock per call); a deep subagent (max_turns=150) emits hundreds of steps on the hot stream loop. Replace with _SubagentEventBuffer, which batches via put_batch (flush on terminal subagent.end, at FLUSH_THRESHOLD, and in the worker finally). 3. build_subagent_step capped only text; tool_calls[].args were copied verbatim, so a large write_file/bash payload produced an unbounded subagent.step row. Cap each call's serialized args at SUBAGENT_STEP_MAX_CHARS, flagged args_truncated. Tests updated/added for all three; AGENTS.md refreshed. * fix(subagents): merge backfill into latest subtask state; reuse message_content_to_text (#3779) Address the remaining two PR review findings: 4. subtask-card's fetchSubtaskSteps().then(updateSubtask) closed over a stale tasks snapshot: a late-resolving backfill wrote setTasks({...stale}), clobbering SSE steps/status and sibling subtasks that arrived during the fetch. useUpdateSubtask now reads/writes through a tasksRef mirroring the latest state (ref-to-latest), and the pure per-subtask transition is extracted to core/tasks/subtask-update.ts::computeNextSubtask (unit-tested). 5. step_events._content_to_text duplicated deerflow.utils.messages. message_content_to_text; call the shared helper instead (guarding None content with 'or ""' so a tool-call-only turn still renders as ""). Tests added for computeNextSubtask and the None-content case; AGENTS.md docs updated.
DeerFlow Frontend
Like the original DeerFlow 1.0, we would love to give the community a minimalistic and easy-to-use web interface with a more modern and flexible architecture.
Tech Stack
- Framework: Next.js 16 with App Router
- UI: React 19, Tailwind CSS 4, Shadcn UI, MagicUI and React Bits
- AI Integration: LangGraph SDK and Vercel AI Elements
Quick Start
Prerequisites
- Node.js 22+
- pnpm 10.26.2+
Installation
# Install dependencies
pnpm install
# Copy environment variables
cp .env.example .env
# Edit .env with your configuration
Development
# Start development server
pnpm dev
# The app will be available at http://localhost:3000
Build & Test
# Type check
pnpm typecheck
# Check formatting
pnpm format
# Apply formatting
pnpm format:write
# Lint
pnpm lint
# Run unit tests
pnpm test
# One-time setup: install Playwright Chromium browser
pnpm exec playwright install chromium
# Run E2E tests (builds and starts production server automatically)
pnpm test:e2e
# Build for production
pnpm build
# Start production server
pnpm start
Site Map
├── / # Landing page
├── /chats # Chat list
├── /chats/new # New chat page
└── /chats/[thread_id] # A specific chat page
Configuration
Environment Variables
Key environment variables (see .env.example for full list):
# Backend API URL (optional, uses local Next.js/nginx proxy by default)
NEXT_PUBLIC_BACKEND_BASE_URL="http://localhost:8001"
# LangGraph-compatible API URL (optional, uses local Next.js/nginx proxy by default)
NEXT_PUBLIC_LANGGRAPH_BASE_URL="http://localhost:8001/api"
Project Structure
tests/
├── e2e/ # E2E tests (Playwright, Chromium, mocked backend)
└── unit/ # Unit tests (mirrors src/ layout)
src/
├── app/ # Next.js App Router pages
│ ├── api/ # API routes
│ ├── workspace/ # Main workspace pages
│ └── mock/ # Mock/demo pages
├── components/ # React components
│ ├── ui/ # Reusable UI components
│ ├── workspace/ # Workspace-specific components
│ ├── landing/ # Landing page components
│ └── ai-elements/ # AI-related UI elements
├── core/ # Core business logic
│ ├── api/ # API client & data fetching
│ ├── artifacts/ # Artifact management
│ ├── config/ # App configuration
│ ├── i18n/ # Internationalization
│ ├── mcp/ # MCP integration
│ ├── messages/ # Message handling
│ ├── models/ # Data models & types
│ ├── settings/ # User settings
│ ├── skills/ # Skills system
│ ├── threads/ # Thread management
│ ├── todos/ # Todo system
│ └── utils/ # Utility functions
├── hooks/ # Custom React hooks
├── lib/ # Shared libraries & utilities
├── server/ # Server-side code
│ └── better-auth/ # Authentication setup and session helpers
└── styles/ # Global styles
Scripts
| Command | Description |
|---|---|
pnpm dev |
Start development server with Turbopack |
pnpm build |
Build for production |
pnpm start |
Start production server |
pnpm test |
Run unit tests with Rstest |
pnpm test:e2e |
Run E2E tests with Playwright |
pnpm format |
Check formatting with Prettier |
pnpm format:write |
Apply formatting with Prettier |
pnpm lint |
Run ESLint |
pnpm lint:fix |
Fix ESLint issues |
pnpm typecheck |
Run TypeScript type checking |
pnpm check |
Run both lint and typecheck |
Development Notes
- Uses pnpm workspaces (see
packageManagerin package.json) - Turbopack enabled by default in development for faster builds
- Environment validation can be skipped with
SKIP_ENV_VALIDATION=1(useful for Docker) - Backend API URLs are optional; nginx proxy is used by default in development
License
MIT License. See LICENSE for details.