Open source direct
Free source (opens in a new tab)
Unknown
For: Developers
Limits
- Model costs are yours
CodingDocumented overviewLast checked Oct 2, 2026
DeerFlow: Pricing, Features & Best Use Cases
Open-source SuperAgent harness for building agents with memory, tools, skills, sandboxes and subagents.
Our take: DeerFlow is for developers comfortable self-hosting an open-source agent harness for research and automation, not for users wanting a hosted service.
Main limitation: DeerFlow must be self-deployed with technical setup, and no hosted offer or hosted pricing was found.
Is DeerFlow right for you?
DeerFlow is for developers comfortable self-hosting an open-source agent harness for research and automation, not for users wanting a hosted service. Based on the vendor's documentation, not on a hands-on test.
Best for: Agent developers
Also a good fit for
- Research teams self-hosting
Consider another tool if must be self-deployed per docs
Or if
- Requires technical setup
- No hosted pricing seen
DeerFlow pricing and usage limits
Prices checked Oct 2, 2026 on the official US pricing page. Prices and limits change often.
Free access and trial conditions
Open source, free to run
None
Paid plans: what changes as you upgrade
Usage caps, credits and extra costs
Model API costs and hosting
Not applicable
What is DeerFlow, and how does it work?
DeerFlow is an open-source framework (code on GitHub under bytedance/deer-flow) for building and running agent systems that research, code and create, handling tasks lasting minutes to hours. Docs describe two parts: a Harness SDK/runtime and a reference App for deployment and end-user workflows.
What you provide and what you get back
You provide
- Research tasks and prompts
- Model provider keys
You get back
- Research reports
- Agent outputs and data visualizations
Where you can use it
- Platforms
- macos, linux
- Account
- Required
- Platforms
- macOS, Linux
- Architecture
- Sandboxes, memory, skills, subagents
- Interface
- Local web frontend
Best use cases for DeerFlow
Scenarios are worked out by our editors from the documentation; the actual result depends on your input and plan.
Deep research reports
- Who
- Researchers
- You provide
- A topic and sources
- Expected result
- You get a researched report produced by agents working on your topic and sources.
- Plan needed
- Free, open source
- Limitation
- Setup and model keys are needed before the harness can run, and costs for the model are yours. source (opens in a new tab)
Build custom agent systems
- Who
- Developers
- You provide
- Skills, tools and subagents
- Expected result
- You get a custom agent system built from your own skills, tools and subagents.
- Plan needed
- Free, open source
- Limitation
- You maintain the deployment yourself, including updates and the environment it runs in. source (opens in a new tab)
Data analysis with visualizations
- Who
- Analysts
- You provide
- A dataset
- Expected result
- You get an analysis of your dataset that includes visualizations and written insights.
- Plan needed
- Free, open source
- Limitation
- The quality of the analysis depends on the model that you choose to connect. source (opens in a new tab)
Key features and their practical limits
Sandboxes
Runs agent work inside sandboxes. Developers who want agent code to run apart from their main system benefit from having that work isolated.
Long-term memory
Keeps memory that persists across sessions. Teams running long tasks benefit because the agent can keep context between sessions.
Tools and skills
Supports tools and skills the agent can call. Developers can extend what an agent is able to do by adding tools and skills it can call.
Subagents
Delegates parts of a task to subagents. People running long tasks benefit because parts of a job can be handed to separate subagents.
DeerFlow: strengths and limits
Where DeerFlow is strong
- Open source Because the code is open source, you can inspect how the agent harness works before running it on your own systems. documented source (opens in a new tab)
- Documentation with tutorials Documentation with tutorials exists, so developers have written guidance to follow when they set the harness up. documented source (opens in a new tab)
- Model provider reference A model provider reference is available, which helps you work out how to configure the model that the agents will use. documented source (opens in a new tab)
- Case studies shown Case studies are shown on the site, so you can see the kinds of tasks the harness is meant to handle before you try it. documented source (opens in a new tab)
Limitations to check before paying
- Must be self-deployed per docs You have to deploy DeerFlow yourself, so you are responsible for running and maintaining the environment it works in. documented source (opens in a new tab)
- Requires technical setup The setup is technical, so people without development experience may not be able to get it running without help. documented source (opens in a new tab)
- No hosted pricing seen No hosted pricing was found, so there is no ready-made paid service to use if you do not want to host it. documented source (opens in a new tab)
- Platform support not itemized Platform support is not itemized in the sources we read, so check that your system works before you plan around it. documented source (opens in a new tab)
How to use DeerFlow for running a research agent locally
Set up DeerFlow
- Open the Quick Start page.
- Install on macOS (brew) or Linux.
- Configure a model provider.
Example: running a research agent locally
- Start the local web frontend.
- Enter a research question.
- Review the generated report.
Review and export the result
- Check that cited sources exist.
- Confirm sandbox runs completed.
Write a report on solid-state battery progress with sources.
Privacy, data use and commercial rights
Summarized from the vendor's current policies; this is not legal advice and a vendor's badge is not an independent audit.
How your inputs and outputs are handled
Runs locally; data goes to the model provider you configure
Depends on your chosen provider
Stored on your machine
Commercial use and plan restrictions
Not stated on the pages read
Not stated on the pages read
Policies: deerflow.tech/privacy
DeerFlow alternatives by need
"Better suited" is based on documented features and pricing, not on a head-to-head test.
Perplexity: better suited to
ChatGPT: better suited to
n8n: better suited to
| Alternative | When to choose it | Pricing | Limitation | |
|---|---|---|---|---|
| Perplexity | — | — | ||
| ChatGPT | — | — | ||
| n8n | — | — |
Explore DeerFlow further
Tools and guides near DeerFlow
More coding tools
Guides and comparisons
DeerFlow user reviews
No user reviews yet.
Used DeerFlow? Be the first to share your experience.
Write a review
DeerFlow FAQ
Is DeerFlow free?
It is open source with no hosted pricing seen; you pay for models and hosting yourself. source (opens in a new tab)
Which platforms are supported?
The Quick Start gives macOS (brew) and Linux setup. source (opens in a new tab)
Does it have memory?
Yes, long-term memory is a listed feature. source (opens in a new tab)
Can it run subagents?
Yes, subagents are a listed feature. source (opens in a new tab)
Sources and review methodology
Official sources and verification dates
- DeerFlow homepage · checked Oct 4, 2026
- DeerFlow Documentation docs · checked Oct 4, 2026
- DeerFlow App Quick Start docs · checked Oct 4, 2026
Not verified: License; Hosted offer; Windows support.
How we documented and maintain this page
Every fact about DeerFlow above comes from the 3 official pages listed here, last compared on Oct 2, 2026; we have not run DeerFlow ourselves, so it has no editorial score. How we check and re-check pages.
Article updated Oct 10, 2026; prices checked Oct 2, 2026. Outbound links to vendors are plain links; we do not earn commissions. How we review · Report an error