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

Arbor

Arbor is an open-source Python autonomous research agent that proposes hypotheses, edits code, runs experiments and keeps improvements that survive held-out-data evaluation.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Arbor — 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 Arbor when a project has a measurable target, an editable codebase and an experiment loop that can distinguish development performance from held-out results.

Who it is for

  • Machine-learning researchers with benchmark-driven optimization tasks.
  • Engineers exploring code, training or data-synthesis changes through repeated experiments.
  • Teams that want experiment outcomes and failed directions retained in a hypothesis tree.

Documented uses

  • Generate and prioritize research hypotheses.
  • Implement code changes for selected ideas.
  • Run experiments in isolated Git worktrees.
  • Compare development results with held-out evaluation data.
  • Track findings from successful and unsuccessful iterations.
Before you choose

Requirements & limits

The repository is public, uses the Apache-2.0 license and documents installation through pip. It also links to a browser-based live demonstration.

  • A benchmark and measurable goal are central inputs to the documented workflow.
  • Experiments are carried out by Executors in isolated Git worktrees.
  • Improvements are evaluated against held-out data before being retained.
  • The project supports multiple model-provider backends through LiteLLM as well as Anthropic and OpenAI interfaces.
  • The documented interface is a CLI; external service integration would require modification.

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

Replay Arbor’s local demonstration

Install Arbor and replay the supplied command-line demonstration of its hypothesis-tree workflow.

Before you start

  • Python package installation with pip.
  • A local command-line environment.

Steps

  1. Install the package with the command below.

    Documented command
    pip install arbor-agent

  2. Run arbor replay --demo to replay the demonstration.

Expected outcome

The documented demo displays the hypothesis tree growing during a local replay.

Guide limitations

  • This recipe covers the supplied replay demonstration, not configuration of a custom research task.
  • The evidence does not establish a service API or external dashboard integration.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
4 October 2026
Original website title
RUC-NLPIR/Arbor · Official homepage title · retained in its source language
How we review sourcesSuggest a correction