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

Optim Agent

Optim Agent is an open-source Python developer tool that uses Claude Code, Codex or OpenCode to propose configurations for systems with declared parameters and measurable objectives. It records evaluated trials and uses parameter context and trial history to guide later proposals.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Optim Agent — 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 your system exposes tunable parameters and each trial can be evaluated with a measurable objective, especially when the parameter meanings and previous results can inform further proposals.

Who it is for

  • Algorithm engineers tuning systems with configurable parameters.
  • Developers who can provide evaluation code and a measurable objective.
  • Researchers running model, inference, quantitative, reinforcement-learning or scientific experiments.

Documented uses

  • Define bounded continuous, integer or categorical parameters for a system.
  • Ask a coding-agent CLI to propose configurations informed by parameter meaning and previous trials.
  • Evaluate configurations against a scalar objective and retain trial outcomes.
  • Review the best recorded parameters and value; optionally request a study summary.
Before you choose

Requirements & limits

The documented setup installs the Python package and requires one authenticated Claude Code, Codex or OpenCode CLI on PATH. The repository identifies the license as MIT.

  • You must provide the objective function and system-specific evaluation; example objectives are not built-in integrations.
  • The coding agent proposes values, while the objective evaluation supplies trial results.
  • Invalid proposals are validated against the declared search space and fall back to safe sampling.
  • Study-wide or per-parameter context can communicate domain meaning to the proposal process.
  • JSON or SQLite storage can retain study data; study summaries are an optional agent-generated feature.

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

Define and run an agent-guided parameter-tuning study

Use a coding-agent CLI to propose parameter values for an objective function and record measured trial results.

Before you start

  • Python package optim-agent installed.
  • One authenticated agent CLI (claude, codex, or opencode) available on PATH.
  • A system with configurable parameters and an objective function that returns a measurable result.

Steps

  1. Define an objective function that declares bounded parameters with trial.suggest_float or trial.suggest_int, optionally adding context to explain their meaning.

  2. Create a study with oa.create_study, setting its direction and an oa.AgentSampler backend matching the available agent CLI.

  3. Call study.optimize(objective, n_trials=20) to evaluate the requested trials.

  4. Inspect study.best_value and study.best_params; optionally set storage to persist the study and summarize=True to request a study summary.

Expected outcome

The study evaluates configurations and exposes the best measured value and corresponding parameters; optional storage preserves study data.

Guide limitations

  • The objective function and domain-specific evaluation must be provided by the user.
  • The agent proposes parameter values; optim-agent validates them against the declared search space, and invalid replies fall back to safe sampling.
  • The documented example uses a finite trial count and does not establish that a particular objective will improve.
Documentation 1 ↗Documentation 2 ↗
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
Optim-Agent/optim-agent · Official homepage title · retained in its source language
How we review sourcesSuggest a correction