theoria

agent
Guvenlik Denetimi
Uyari
Health Uyari
  • No license — Repository has no license file
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 10 GitHub stars
Code Gecti
  • Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Gecti
  • Permissions — No dangerous permissions requested

Bu listing icin henuz AI raporu yok.

SUMMARY

Agent gauge platform, comparison, operational observation.

README.md
Theoria

Theoria

English | 简体中文

Put every AI agent on the same starting line

A local-first desktop workspace for parallel AI coding agents and reproducible evaluations

Privacy first. All data stays local, with application records stored in SQLite.

Tauri 2 React 19 TypeScript 5.8 Rust Stable Desktop

Product Preview · Highlights


Product Preview

Parallel agent runs in Theoria
Run multiple agents on the same task and follow their status, responses, and tool calls in real time
Skill management in Theoria
Manage skills centrally and mount them into workspaces
Benchmark management in Theoria
Organize reproducible benchmark comparisons

About Theoria

Theoria is a desktop workspace for AI coding agents. It gives every selected agent the same workspace snapshot and an isolated execution directory, allowing Codex, Claude Code, OpenCode, and WorkBuddy to solve one task in parallel while their progress, results, and file changes remain easy to compare.

It supports everyday multi-agent development workflows and provides consistent environments, run history, and skill configuration for repeatable capability evaluations.

[!WARNING]
This project is not yet complete and remains under active development.

Highlights

Capability Description
Parallel agents Run up to six agents on one task, including Codex, Claude Code, OpenCode, and WorkBuddy
Isolated execution Start from an immutable workspace snapshot and give each agent its own working directory
Observable progress Follow status, streaming output, tool calls, token usage, and duration; stop one agent or all of them
Continued collaboration Preserve agent sessions and send follow-up prompts to every agent or a selected subset
Result comparison Collect final responses and file changes, then persist run history for later review
Skill management Manage skills from local or Git sources and mount them into one or more workspaces
Benchmark workflows Organize reproducible comparisons around fixed snapshots, test cases, and selected agents
Local-first data Keep workspaces on the local file system and persist application records in SQLite
Internationalization Use the built-in Simplified Chinese and English interfaces with localized frontend and backend errors

Yorumlar (0)

Sonuc bulunamadi