OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Research Agents

Autoact

Autoact is an open-source Python research framework for learning question-answering agents through self-planning, tool selection and synthesized planning trajectories.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Autoact — OpenViST project profile in Research Agents.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Autoact to inspect or experiment with a research workflow for training QA agents from limited task data and synthesized plans, rather than as a hosted question-answering service.

Who it is for

  • Researchers studying language-agent learning and self-planning.
  • Developers examining or adapting Python code for question-answering agents and automatic tool selection.

Documented uses

  • Generate task data through the documented self-instruct workflow.
  • Select applicable tools for a task using a tool library.
  • Synthesize planning trajectories for question-answering tasks.
  • Train differentiated planning, tool-use and reflection agents from trajectories.
Before you choose

Requirements & limits

The repository is public and lists an Apache-2.0 license in repository metadata. The README displays an MIT license badge, so check the repository license file before reuse. Its experiment instructions require a Bing Search key described as not free.

  • The workflow depends on a tool library and, for experiments, a Bing Search key.
  • The README’s training example uses DeepSpeed, LoRA, a Llama 2 chat model and eight CUDA device IDs.
  • The README describes training code adapted from FastChat, inference based on BOLAA, and LangChain use with open models via FastChat.
  • The repository metadata lists Apache-2.0, while its README badge says MIT; verify the license file.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install the Autoact repository and prepare for experiments

Set up the documented repository environment and identify the experiment prerequisite.

Before you start

  • Git
  • Python and pip
  • A Bing Search key for experiments; the README says this key is not free.

Steps

  1. Clone https://github.com/zjunlp/AutoAct and change into the AutoAct directory.

  2. Install the listed dependencies with pip install -r requirements.txt.

  3. Apply for a Bing Search key before running experiments.

Expected outcome

The repository dependencies are installed, and the documented search-key prerequisite is identified for experiments.

Guide limitations

  • The installation instructions do not document a specific Python version or operating-system setup.
  • This covers setup only, not running an experiment or confirming compatibility.
  • The project lists Apache-2.0 in repository metadata, while the README badge displays MIT; check the repository license file for reuse terms.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
zjunlp/AutoAct · Official homepage title · retained in its source language
How we review sourcesSuggest a correction