Rockachopa/Timmy-time-dashboard
1
0
Fork 2
Code Issues 7 Pull Requests Actions Packages Projects Releases Wiki Activity

research: evaluate DeerFlow as autonomous research orchestration layer #1283

New Issue
Closed
opened 2026-03-24 01:44:06 +00:00 by claude · 1 comment
claude commented 2026-03-24 01:44:06 +00:00
Collaborator
Copy Link

Context

From screenshot triage (issue #1275).

DeerFlow (deer-flow) is an autonomous research pipeline framework with a Docker-based setup. It runs at localhost:2026 and appears to support multi-agent orchestration for tasks like market research, code analysis, and data gathering.

Repo: https://github.com/bytedance/deer-flow (ByteDance)

Problem

The Timmy autonomous research pipeline (#972) was implemented as a custom orchestrator in SOVEREIGNTY.md. Before investing further in the custom solution, it is worth evaluating DeerFlow to see if it provides capabilities we can integrate or learn from.

Research Questions

  1. What agent roles does DeerFlow define? How do they map to Timmy’s personas?
  2. Does DeerFlow expose an API we can call from Timmy’s research pipeline?
  3. What LLM backends does DeerFlow support? Can it use Ollama/vLLM?
  4. Is the license compatible with Timmy’s use case?
  5. Does the Docker setup conflict with our existing stack on any ports?

Investigation Tasks

  • Clone and run DeerFlow locally (make docker-init && make docker-start)
  • Document the agent architecture (roles, tools, handoff protocol)
  • Test with Ollama as the LLM backend
  • Assess integration options: full adoption vs selective borrowing vs none
  • Write a short findings note and link it from this issue

Acceptance Criteria

A comment on this issue with:

  • Go / No-go recommendation for integration
  • If Go: a follow-up implementation issue with a concrete plan
  • If No-go: which specific patterns from DeerFlow are worth borrowing

References

  • Issue #972 — autonomous research pipeline already in Timmy
  • IMG_6124 screenshot: Docker Compose setup, provisioner mode with config.yaml
## Context From screenshot triage (issue #1275). DeerFlow (`deer-flow`) is an autonomous research pipeline framework with a Docker-based setup. It runs at `localhost:2026` and appears to support multi-agent orchestration for tasks like market research, code analysis, and data gathering. **Repo:** https://github.com/bytedance/deer-flow (ByteDance) ## Problem The Timmy autonomous research pipeline (#972) was implemented as a custom orchestrator in SOVEREIGNTY.md. Before investing further in the custom solution, it is worth evaluating DeerFlow to see if it provides capabilities we can integrate or learn from. ## Research Questions 1. What agent roles does DeerFlow define? How do they map to Timmy’s personas? 2. Does DeerFlow expose an API we can call from Timmy’s research pipeline? 3. What LLM backends does DeerFlow support? Can it use Ollama/vLLM? 4. Is the license compatible with Timmy’s use case? 5. Does the Docker setup conflict with our existing stack on any ports? ## Investigation Tasks - [ ] Clone and run DeerFlow locally (`make docker-init && make docker-start`) - [ ] Document the agent architecture (roles, tools, handoff protocol) - [ ] Test with Ollama as the LLM backend - [ ] Assess integration options: full adoption vs selective borrowing vs none - [ ] Write a short findings note and link it from this issue ## Acceptance Criteria A comment on this issue with: - Go / No-go recommendation for integration - If Go: a follow-up implementation issue with a concrete plan - If No-go: which specific patterns from DeerFlow are worth borrowing ## References - Issue #972 — autonomous research pipeline already in Timmy - IMG_6124 screenshot: Docker Compose setup, provisioner mode with `config.yaml`
claude was assigned by Timmy 2026-03-24 01:46:44 +00:00
claude was unassigned by Timmy 2026-03-24 01:56:15 +00:00
claude commented 2026-03-24 01:56:29 +00:00
Author
Collaborator
Copy Link

Research complete. PR #1305 adds docs/research/deerflow-evaluation.md.

Go / No-go: No-go for full adoption

Timmy already has a working ResearchOrchestrator (#975). DeerFlow is a full-stack harness (Python 3.12+ + Node.js + LangGraph + Nginx) — the lock-in cost doesn't justify displacing the custom solution at this stage.

Key findings

Question Answer
Agent roles Lead agent + dynamic sub-agents (LangGraph); 12-middleware chain; 3-concurrent fan-out
REST API Full FastAPI + LangGraph server, no auth; callable from Timmy via httpx
LLM backends OpenAI-compatible base_url override — Ollama works (community-confirmed: Qwen2.5-14B, Llama 3.1)
License MIT — fully compatible
Port conflicts None — DeerFlow uses port 2026; Timmy uses 8000/8080/11434

Patterns worth borrowing

  1. Parallelize ResearchOrchestrator with asyncio.gather() — DeerFlow fans out 3 sub-agents concurrently; Timmy is currently sequential. Highest ROI, no new dependencies.
  2. Context-trimming step before synthesis (mirrors DeerFlow's SummarizationMiddleware — prevents token budget blowout on long research runs)
  3. MCP server discovery in research_tools.py for pluggable tool backends
  4. Revisit DeerFlow's code-execution sandbox if #978 (Paperclip task runner) proves insufficient

Recommended follow-up issues

  • Parallelize ResearchOrchestrator query execution (asyncio.gather) — medium priority
  • Add context-trimming to synthesis cascade — low priority
  • MCP server discovery in research_tools.py — low priority
  • Re-evaluate after #980 + #981 land to confirm custom orchestrator is tracking sovereignty targets
Research complete. PR #1305 adds `docs/research/deerflow-evaluation.md`. --- ## Go / No-go: **No-go for full adoption** Timmy already has a working `ResearchOrchestrator` (#975). DeerFlow is a full-stack harness (Python 3.12+ + Node.js + LangGraph + Nginx) — the lock-in cost doesn't justify displacing the custom solution at this stage. ## Key findings | Question | Answer | |----------|--------| | Agent roles | Lead agent + dynamic sub-agents (LangGraph); 12-middleware chain; 3-concurrent fan-out | | REST API | Full FastAPI + LangGraph server, no auth; callable from Timmy via httpx | | LLM backends | OpenAI-compatible base_url override — Ollama works (community-confirmed: Qwen2.5-14B, Llama 3.1) | | License | **MIT** — fully compatible | | Port conflicts | **None** — DeerFlow uses port 2026; Timmy uses 8000/8080/11434 | ## Patterns worth borrowing 1. **Parallelize ResearchOrchestrator** with `asyncio.gather()` — DeerFlow fans out 3 sub-agents concurrently; Timmy is currently sequential. Highest ROI, no new dependencies. 2. **Context-trimming step** before synthesis (mirrors DeerFlow's SummarizationMiddleware — prevents token budget blowout on long research runs) 3. **MCP server discovery** in `research_tools.py` for pluggable tool backends 4. Revisit DeerFlow's code-execution sandbox if #978 (Paperclip task runner) proves insufficient ## Recommended follow-up issues - Parallelize ResearchOrchestrator query execution (`asyncio.gather`) — medium priority - Add context-trimming to synthesis cascade — low priority - MCP server discovery in `research_tools.py` — low priority - Re-evaluate after #980 + #981 land to confirm custom orchestrator is tracking sovereignty targets
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
222-epic
Workshop: Timmy as Presence (Epic #222)
 actionable
Has a concrete code/config task extracted
 assigned-claude
Issue currently assigned to Claude agent — do not assign to another agent
 assigned-gemini
Issue currently assigned to Gemini agent — do not assign to another agent
 assigned-groq
assigned-kimi
Issue currently assigned to Kimi agent — do not assign to another agent
 assigned-manus
Issue currently assigned to Manus agent — do not assign to another agent
 claude-ready
consolidation
Part of a consolidation epic
 deprioritized
Keep open but not blocking P0 work
 deprioritized
Keep open but not blocking P0 work
 duplicate
Duplicate of another issue
 gemini-review
Auto-generated by Gemini, needs relevance review
 groq-ready
harness
Core product: agent framework, heartbeat, inference, memory
 heartbeat
Harness: Agent heartbeat loop
 inference
Harness: Inference and model routing
 infrastructure
Supporting stage: dashboard, CI/CD, deployment, DNS
 kimi-ready
Scoped and ready for Kimi to pick up
 memory-session
Harness: Memory and session crystallization
 morrowind
Harness: Morrowind embodiment
 needs-design
Needs architectural design before implementation
 needs-extraction
Philosophy with unextracted engineering work
 p0-critical
Priority 0: Must fix now
 p1-important
Priority 1: Important, next sprint
 p2-backlog
Priority 2: Backlog, do when time permits
 philosophy
Philosophical foundation — informs architecture decisions
 rejected-direction
Closed: rejected or superseded direction
 seed:know-purpose
Three Seeds: KNOW YOUR PURPOSE
 seed:serve-real
Three Seeds: SERVE THE REAL
 seed:tell-truth
Three Seeds: TELL THE TRUTH
 sovereignty
Harness: Sovereignty stack
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
Timmy claude grok groq kimi manus perplexity Rockachopa
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Rockachopa/Timmy-time-dashboard#1283
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?