ProAgent Workflow Automation
Turns human instructions into n8n workflows and coordinates specialized agents for workflow decisions.
- Source repo
- OpenBMB/ProAgent
- Stars
- ★ 864
- Last updated
- 2y ago
- License
- Apache-2.0
- Primary language
- Python
- FA score
- 45/100 · Major gaps
At a glance
- How it runs
- Works with
- Platform-specificOpenAI API (Partial support)
- Cost
- Free software; you pay for model usage
- Setup effort
- High · needs real infrastructure
- You'll need
- Typical use
- An operations team with self-hosted n8n that wants to construct cross-application workflows from natural-language requests.
- Not a fit if
- Users wanting a turnkey tool without setting up n8n and registering third-party apps
- Teams on the latest n8n or newest OpenAI API; the compiler and interface target older versions
- Source review
- 45/100 · Major gaps 1 safety controls not found
What does this agent do, and when should you use it?
ProAgent is the official code for the Agentic Process Automation (APA) paper, intended to turn human instructions into automation workflows. It operates around self-hosted n8n and expects n8n-exported workflow and credential files to be supplied to the codebase. Runtime behavior is configured in ProAgent/config.py: development builds workflows as described in the paper, refine modifies an existing workflow, and production loads an existing run for reproduction. Each run creates a readable record under ./records that refine and production can load later. Development requires OpenAI API configuration; the repository states that the implementation is based on GPT4-0613 and an older OpenAI interface version.
After a runtime mode is selected in ProAgent/config.py, the entry point is python main.py. In development mode, ProAgent builds workflows from human instructions, coordinates specialized agents for complex decisions, and can ask a human for help through a function call; n8n must be started first. The code consumes c.json and w.json exported through n8n export:credentials and n8n export:workflow, then writes run records to ./records. Refine loads an existing workflow and changes it for a new request with test-on-change enabled; production loads an existing run to reproduce it and tests APA-code only once at the end.
- An operations team with self-hosted n8n that wants to construct cross-application workflows from natural-language requests.
- A researcher reproducing the reported paper cases by loading an existing run from ./apa_case in production mode.
- An automation maintainer who has an existing n8n workflow and needs to refine it for a new request.
- A team building and testing workflows with human intervention when the system needs help understanding a request.
- A developer who has registered applications and credentials in n8n and needs the code to load workflow and credential IDs.
How do you install or deploy this agent?
Install Python packages:
pip install -r requirements.txtTo connect to real application services, install and run self-hosted n8n:
npm install n8n -g
export WEBHOOK_URL=https://n8n.x-agent.net/redirect/http%3A%2F%2Flocalhost%3A5678/
n8nAfter connecting or registering apps in n8n, export credentials and workflows:
n8n export:credentials --all --decrypted --output=./ProAgent/n8n_tester/credentials/c.json
n8n export:workflow --all --output=./ProAgent/n8n_tester/credentials/w.jsonDevelopment mode additionally requires OPENAI_API_KEY and OPENAI_API_BASE.
How do you use this agent?
Select development, refine, or production in ProAgent/config.py. For development, start n8n and set OPENAI_API_KEY plus OPENAI_API_BASE, then run:
python main.pyTo reproduce a saved case, use production mode to load an existing run from ./apa_case or ./records; this reproduction path does not require n8n. Use refine mode to load an existing workflow and improve it for a new request.
What are this agent's strengths and limitations?
- Combines workflow construction and execution-time decisions in one APA process instead of only running pre-authored RPA steps.
- Provides distinct development, refine, and production modes for construction, iterative changes, and run reproduction.
- Writes readable records under ./records that can be reloaded by refine or production.
- Includes a function-call path for proactively requesting human help when the request has been misunderstood.
- Core operation depends on OpenAI configuration and is explicitly based on GPT4-0613 plus an older OpenAI interface version, creating version-compatibility risk.
- Connecting real applications requires self-hosted n8n, pre-registered apps, and exported decrypted credentials and workflows.
- The repository states that its n8n compiler targets an older n8n version and may not work with newer releases.
- Self-hosted n8n setup and third-party account connections can encounter unaddressed problems; the repository also notes possible network restrictions in China.
How does this agent compare with similar options?
Against traditional RPA, ProAgent is positioned for workflow design and dynamic execution decisions that need human-like intelligence. Rather than only following fixed automation steps, it attempts to construct workflows from human instructions and coordinate specialized agents for decisions.
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| ProAgent Workflow Automation This agent | 45 · Major gaps | CLIFree + model costs | ★ 864 | 2y ago | Python | — |
| superglue Enterprise Integration Agents | 45 · Major gaps | Hosted serviceFreemium | ★ 2.1k | 12d ago | TypeScript | OpenAI API · Claude API |
| OpenClaw Chinese Edition | 53 · Major gaps | CLIFree + model costs | ★ 3.8k | today | JavaScript | OpenAI API · Claude API |
| LocalAI – Open-Source AI Engine | 52 · Major gaps | Self-hosted serviceFree | ★ 50k | today | Go | OpenAI API · Claude API |
How does FollowAgents rate this agent?
Why each dimension lost points
The README says ProAgent can connect to self-hosted n8n and real apps; main.py's example reads Google Sheets and sends profit information through Slack or email, indicating broad external permissions and effects. The packet shows no least-privilege setup or sensitive-data safeguards, so those scores are low. The README describes human collaboration and a function call for help, but does not establish pre-execution confirmation or rollback; the run-record feature offers only limited recovery evidence. The paper link and citation support attribution, while publisher identity remains unknown.
Dependencies are pinned. The README says an older OpenAI API version is required and warns that the n8n compiler may be incompatible with newer n8n and that the redirect service may be unstable, providing some consistency and dependency-risk disclosure. The packet gives no concrete failure-handling or error-message policy, so failure messages receive only partial credit.
The README describes development, refine, and production modes, including paper reproduction and refinement of existing workflows; the code example also gives a specific business process. It discloses an older n8n compatibility boundary and production-mode testing timing, but the trigger is mainly illustrated as a manual trigger, and environment setup still depends on users configuring n8n, accounts, and API credentials.
The README organizes setup, n8n preparation, operating modes, and citation, with commands, credential-export steps, and a reproduction case; it also discloses some compatibility and service-stability limits. The LICENSE contains the Apache-2.0 text. Naming and configuration descriptions include spelling and wording inconsistencies; version-change information is limited, and maintenance responsibility and update path are unclear, so these areas do not receive full marks.
The stated goal is to turn natural-language requests into workflows and coordinate specialized agents. The example combines spreadsheet calculations, business classification, and notifications, which could reduce manual workflow construction. The packet describes run records and workflow reuse, but also requires n8n, external accounts, API keys, and environment maintenance, so cost-benefit evidence is partial.
The README ties the project claim to a paper and identifies reproduction modes, a case directory, and configuration location; main.py provides a concrete workflow task that partially aligns with the product description. However, this packet contains only the README, entry point, dependency list, and license, without the paper contents, implementation details, or multiple independent materials to verify core capabilities. Descriptive claims are not always clearly separated from demonstrated facts.
- Not found in source: sensitive-data handlingUse dedicated, low-privilege, revocable API keys — never production credentials — and keep secrets out of logs.
- The example reads business cost and sales data and sends results through Slack or email; verify n8n credential scope, recipients, and data handling before deployment.
- The README instructs users to export all n8n credentials in decrypted form to a file; that file may contain sensitive information, and the packet does not describe encryption, access controls, or secure deletion.
- The README says the redirect service may be unstable and the compiler may be incompatible with newer n8n versions; it states that users may need to handle environment issues themselves.
Maintain this project? Add the FARS badge to your README
[](https://followagents.com/en/agents/proagent)<a href="https://followagents.com/en/agents/proagent"><img src="https://followagents.com/badge/proagent.svg" alt="FARS review"></a>FAQ
Does it require OpenAI?
OPENAI_API_KEY and OPENAI_API_BASE. The repository identifies GPT4-0613 and an older OpenAI interface version, and does not document support for other model providers.Is n8n always required?
How does it obtain access to third-party apps?
./ProAgent/n8n_tester/credentials/c.json, alongside the exported workflow file.Will it work unchanged with current n8n releases?
Thanks — we use this to improve the page.