Back to results

Agent · Multi-step assistant · bytedance / contributors

DeerFlow · Research and creation agent

An open-source agent environment combining tools, skills and multi-step workflows for research, analysis and content creation.

Agent source setupDeveloper integrationDeveloper knowledge neededUsage based

Opens the provider’s website. Check its account, setup and pricing requirements below.

Use it inLocal environment · Docker · Web browser · Python environment
Account & accessAPI key required
Cost to expect

Source is available under the listed license; models, cloud hosting and external services may cost extra.

Available from the publisher

This resource is not hosted as a local package

Use the source code and setup instructions from the publisher. A complete local source package is not available here.

View original source
From knowing to doing

A clear path to your first result

01
Prepare these firstEnvironment, account and materials
  • Prepare Docker or the documented local runtime and a model configuration; web research needs a working search service.
02
Follow the setup stepsUse the guide for this specific resource
  1. 1
    Prepare Docker or the documented local runtime and a model configuration; web research needs a working search service.
  2. 2
    Follow the official quickstart and configuration wizard to select model, search and execution settings, then start the local workspace.
  3. 3
    Start with a test task without sensitive data, inspect execution records and results, then decide whether to use it for real work.
03
Adapt a real task exampleA starter prompt for the compatible app

Research an industry topic, first propose a plan, then deliver a cited report and unresolved questions.

Paste this task in the resource’s compatible app or host after setup. Replace the context with your own material, then check the result.

04
Check the result before using itLimits, permissions and your first-run notes

Limits & permissions

  • This site provides navigation and source links. It does not host this agent in your browser; installation and model setup are required.
  • The official README and license were reviewed; local installation and end-to-end execution have not been tested.

Access involved: Use network services · Run configured tool code

Your first-run checklist

Catalog review covers published metadata. It does not certify that the resource runs in your environment. These checks are your own local notes and reset for a new catalog revision.

This is a Markdown checklist of steps and links. It is not an Agent or Skill installation package.

GitHub source & original repositorybytedance/deer-flow · Publisher files and version details

Get it from the source repository

This repository provides Agent source or configuration. Check its setup instructions to understand the environment and accounts it needs.

Source version
Linked version · fcbb14e83f1c

The version ZIP contains the entire repository at the linked commit, including other resources. Find the folder for this resource before using it.

  1. 1

    Read the matching source

    Open the resource location and read its README or SKILL.md. Check the task, supported client and prerequisites.

  2. 2

    Get files on GitHub

    Use the linked-version repository ZIP to match this guide, or open the repository home page to check its current content. A source ZIP is different from a ready-to-run installer.

  3. 3

    Follow the publisher’s setup

    Prepare the documented environment and required accounts. Start with the provided example and check the result before using it for your work.

Read the exact linked source
More AI resource navigation
AI resource index and instructions

Get it from the source repository

This repository provides Agent source or configuration. Check its setup instructions to understand the environment and accounts it needs.

Repository: bytedance/deer-flow

Linked version · fcbb14e83f1c

The version ZIP contains the entire repository at the linked commit, including other resources. Find the folder for this resource before using it.

  1. Read the matching source

    Open the resource location and read its README or SKILL.md. Check the task, supported client and prerequisites.

  2. Get files on GitHub

    Use the linked-version repository ZIP to match this guide, or open the repository home page to check its current content. A source ZIP is different from a ready-to-run installer.

  3. Follow the publisher’s setup

    Prepare the documented environment and required accounts. Start with the provided example and check the result before using it for your work.

Requirements and use guide

  • Prepare Docker or the documented local runtime and a model configuration; web research needs a working search service.
  1. Prepare Docker or the documented local runtime and a model configuration; web research needs a working search service.
  2. Follow the official quickstart and configuration wizard to select model, search and execution settings, then start the local workspace.
  3. Start with a test task without sensitive data, inspect execution records and results, then decide whether to use it for real work.

Task example

Research an industry topic, first propose a plan, then deliver a cited report and unresolved questions.

Limits

  • This site provides navigation and source links. It does not host this agent in your browser; installation and model setup are required.
  • The official README and license were reviewed; local installation and end-to-end execution have not been tested.

Explore by task