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

Codex Autoresearch

Codex Autoresearch is a TypeScript Codex plugin that runs bounded, repeatable optimization experiments against local code. It combines a workload, correctness checks, an edit boundary and a time budget to record a baseline, evaluate small changes and leave a reviewable patch with evidence.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Codex Autoresearch — OpenViST project profile in Coding Agents.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when the desired outcome can be expressed as a repeatable metric or explicit conditions, and when the allowed files, checks and experiment budget can be agreed in advance.

Who it is for

  • Developers optimizing code against a repeatable benchmark.
  • Engineering teams with measurable performance or resource-use targets.
  • Codex users who want explicit experiment contracts and bounded change scope.

Documented uses

  • Reduce test or build runtime against a defined workload.
  • Investigate memory, throughput, bundle-size or query-latency changes.
  • Compare candidate code changes while preserving independent correctness checks.
  • Prepare an evidence-backed patch for review.
Before you choose

Requirements & limits

The repository is public, open source and licensed under Apache-2.0. It is described as usable with cloud or self-hosted Codex workflows.

  • A repeatable evaluator is required for the standard benchmark path.
  • The benchmark must print a primary metric such as METRIC seconds=12.34.
  • Correctness checks and secondary metrics can define acceptable trade-offs.
  • The contract specifies scope, keep and stop rules, noise and budgets before execution.
  • The tool is implemented in TypeScript and is built around the Codex ecosystem.

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

Set up a bounded optimization experiment

Prepare and approve a repeatable Codex experiment for a measurable code target.

Before you start

  • A Codex installation with the plugin picker available.
  • A repository containing the workload, correctness checks and permitted edit scope.
  • A benchmark that prints a primary metric, plus a defined time or packet budget.

Steps

  1. Open the Codex plugin picker with /plugins.

  2. Choose TheGreenCedar -> codex-autoresearch -> Install plugin, then start a new Codex task in the target repository.

  3. Provide the goal, benchmark, metric direction, checks, scope, protected paths and stop budget.

  4. Review and accept the complete experiment contract before setup or execution.

  5. Allow the workflow to measure the baseline and proceed through the documented experiment loop.

Expected outcome

An accepted experiment contract, baseline measurement and a logged, reviewable optimization result when useful work is available.

Guide limitations

  • The benchmark and paths in the example must be adapted to the project.
  • The repository states that no model-driven comparison has been run against ordinary Codex or version 2.9.0.
  • The guide describes the documented setup and workflow; it does not establish successful performance improvement.
Documentation 1 ↗
Evidence and transparency

Sources & review

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