DeerFlow
ByteDance's open-source super-agent harness for deep research, reports and multi-step tasks with sandboxes.
Quick answer
DeerFlow is a byteDance's open-source super-agent harness for deep research, reports and multi-step tasks with sandboxes, listed under AI search & research assistants. A typical small-business use: market and competitor research reports with sources, generated on a schedule. Official hardware: minimum 8 vCPU · 16 GB RAM · 40 GB disk (recommended 16 vCPU · 32 GB RAM). That fits a Hetzner Cloud CPX42 server at about €69.99/month (server only, excl. VAT; €130.49/month for the recommended spec). License: MIT, which allows internal business use.
Last updated · facts verified against official docs on · sources
Activity data comes from GitHub and was last refreshed Oct 28, 2026.
What is DeerFlow?
DeerFlow 2.0 is a ground-up rewrite of ByteDance's deep research framework into a general agent harness. A lead agent plans and delegates to sub-agents with their own tools and context, runs code and files in sandbox containers, keeps long-term memory, and can browse, search the web or query a private RAGFlow knowledge base. It has a web UI, accounts with admin roles, scheduled tasks, chat channel bots (Slack, Feishu/Lark, WeCom, DingTalk, Telegram) and works with many LLM providers.
Where it fits in a 10–200 person business
- Market and competitor research reports with sources, generated on a schedule
- An internal research assistant that combines web search with company documents in RAGFlow
- Multi-step analysis tasks that need code execution in an isolated sandbox
These use cases are our editorial suggestions, not claims made by the project.
Can a business use DeerFlow? (license)
| License | MIT |
|---|---|
| Type | Permissive: MIT / Apache-2.0: use, modify and sell with few obligations. |
| Internal business use | Allowed under the license text |
| What to watch | Permissive. Internal and commercial use allowed; keep the license notice. |
This is a plain-English summary of the license text, not legal advice. Third-party models you run with it have their own licenses.
Data & compliance: does DeerFlow phone home?
| Telemetry (usage data sent to the vendor) | Not documented |
|---|---|
| Other calls home by default (updates, marketplaces, default cloud services) | Not documented |
| Single sign-on (OIDC / SAML / LDAP) | Not documented |
| Audit log | Not documented |
| Role-based access control (RBAC) | Yes, free Pluggable authorization (authorization.enabled in config.yaml) can deny tools, models, sandbox execution and run routes per role; it is off by default. |
| Data residency / GDPR notes from the project | Not documented |
Checked against the project's own docs and repository on . "Not documented" means we could not find the answer in official sources; it does not mean the feature is missing or that no data is sent. Third-party services you connect (model APIs, vector databases, search engines) have their own data policies. Compare all apps.
What hardware does DeerFlow need?
| Minimum | 8 vCPU · 16 GB RAM · 40 GB disk |
|---|---|
| Recommended | 16 vCPU · 32 GB RAM |
Figures are the README's sizing table for a long-running server (make up): "starting point" 8 vCPU, 16 GB RAM, 40 GB free SSD; recommended 16 vCPU, 32 GB RAM. Local evaluation starts at 4 vCPU, 8 GB RAM, 20 GB SSD ("2 vCPU / 4 GB is usually not enough"). The numbers exclude any local LLM.
How much does it cost to host DeerFlow per month?
| Minimum spec | 8 vCPU · 16 GB RAM · 40 GB disk → CPX42 (8 vCPU, 16 GB, 320 GB disk) €69.99/mo |
|---|---|
| Recommended spec | 16 vCPU · 32 GB RAM → CPX62 (16 vCPU, 32 GB, 640 GB disk) €130.49/mo |
How we got this: we took the cheapest Hetzner Cloud Regular Performance (shared AMD vCPU) plan with at least the RAM and vCPU the project lists, using prices read from www.hetzner.com/cloud/regular-performance on 2026-10-25. "From" monthly price cap shown by Hetzner, excluding VAT. Prices vary by location. Covers the server only; it excludes LLM API usage, backups, extra storage, GPUs and your time. Disk is not matched; add a volume if you need more than the plan includes.
How to deploy DeerFlow
Method: Docker Compose (via make)
git clone https://github.com/bytedance/deer-flow.git
cd deer-flow
make setup # wizard: LLM provider, web search, sandbox; writes config.yaml and .env
make up # build images and start production services
# open http://localhost:2026 and create the admin account at /setupFrom the official instructions: github.com/bytedance/deer-flow#running-the-application. Always check the source for the current version before you run it.
Known gotchas when self-hosting DeerFlow
- DeerFlow can execute system commands and is designed for a trusted local environment reachable only on 127.0.0.1. Before exposing it (BIND_HOST), add an IP allowlist and an authenticating reverse proxy. source
- A fresh instance has no accounts; create the admin at /setup immediately after starting any deployment that is not loopback-only. Gateway admin is equivalent to code execution on the host. source
- Docker deployment requires Docker Compose v2.24 or newer; older clients cannot parse the compose files. source
- Version 2.0 shares no code with 1.x; guides written for the original Deep Research framework apply to the main-1.x branch only. source
Alternatives in AI search & research assistants
- Khoj Personal and team AI "second brain" for your documents and the web. Compare DeerFlow vs Khoj
- Vane (formerly Perplexica) Private, Perplexity-style AI answer engine with cited web sources. Compare DeerFlow vs Vane (formerly Perplexica)
- GPT Researcher Open-source deep research agent that browses the web and local documents and writes cited reports. Compare DeerFlow vs GPT Researcher
- SurfSense Open-source NotebookLM / Perplexity-style research workspace you can self-host. Compare DeerFlow vs SurfSense
DeerFlow FAQ
How much RAM does DeerFlow need?
The official minimum is 8 vCPU · 16 GB RAM · 40 GB disk. The recommended spec is 16 vCPU · 32 GB RAM. Figures are the README's sizing table for a long-running server (make up): "starting point" 8 vCPU, 16 GB RAM, 40 GB free SSD; recommended 16 vCPU, 32 GB RAM. Local evaluation starts at 4 vCPU, 8 GB RAM, 20 GB SSD ("2 vCPU / 4 GB is usually not enough"). The numbers exclude any local LLM.
How much does it cost to self-host DeerFlow?
The software is free to download. Based on the official minimum 8 vCPU · 16 GB RAM · 40 GB disk (recommended 16 vCPU · 32 GB RAM): That fits a Hetzner Cloud CPX42 server at about €69.99/month (server only, excl. VAT; €130.49/month for the recommended spec). That covers the server only, not LLM API usage, backups, extra storage or your time.
Is DeerFlow free for commercial use?
DeerFlow is licensed under MIT (Permissive). Internal business use is allowed under the license text. Permissive. Internal and commercial use allowed; keep the license notice. This is a summary, not legal advice.
What are common problems when self-hosting DeerFlow?
DeerFlow can execute system commands and is designed for a trusted local environment reachable only on 127.0.0.1. Before exposing it (BIND_HOST), add an IP allowlist and an authenticating reverse proxy. A fresh instance has no accounts; create the admin at /setup immediately after starting any deployment that is not loopback-only. Gateway admin is equivalent to code execution on the host. Docker deployment requires Docker Compose v2.24 or newer; older clients cannot parse the compose files. Version 2.0 shares no code with 1.x; guides written for the original Deep Research framework apply to the main-1.x branch only.
Is DeerFlow still maintained?
Yes, it is actively developed. Last commit: Oct 28, 2026. Latest release: v2.1.0 (Sep 24, 2026). Checked Oct 28, 2026.
How do I install DeerFlow?
The official method is Docker Compose (via make). The exact commands are in the "How to deploy" section above and in the official instructions: https://github.com/bytedance/deer-flow#running-the-application
What are self-hosted alternatives to DeerFlow?
Other apps in AI search & research assistants: Khoj, Vane (formerly Perplexica), GPT Researcher, SurfSense.
Sources
- GitHub repository: github.com/bytedance/deer-flow
- Official documentation: github.com/bytedance/deer-flow#quick-start
- License text: github.com/bytedance/deer-flow/blob/main/LICENSE
- Hardware requirements: github.com/bytedance/deer-flow#deployment-sizing
- Deploy instructions: github.com/bytedance/deer-flow#running-the-application
- Gotcha 1: github.com/bytedance/deer-flow#%EF%B8%8F-security-notice
- Gotcha 2: github.com/bytedance/deer-flow#deployment-defaults
- Gotcha 3: github.com/bytedance/deer-flow#option-1-docker-recommended
- Gotcha 4: github.com/bytedance/deer-flow#-deerflow---20
- Hetzner Cloud pricing (cost estimate): www.hetzner.com/cloud/regular-performance
- GitHub stars, commits and releases: GitHub REST API, fetched
Facts on this page were checked against the project's own repository and docs on 2026-10-28. Spotted something out of date? Tell us.