Myra Agents
macOS · Windows · Linux · Web Open source

Control your AI agents from anywhere

It's never been easier to run, manage, and monitor your AI agents. Spawn proprietary or open-source agents from anywhere, on all your devices. Run a dozen at once. Schedule them to recur.

Myra Agents Kanban board with Draft, To Do, In Progress and Waiting lanes

One board for every agent you run

A board, a queue, and a scheduler wrapped around the AI agent you already use.

See where every task stands

A native Kanban view of your agents. Cards move themselves: Draft → To Do → In Progress → Waiting Feedback → Awaiting Review → Done, driven by what the agent reports.

Bring your own agent

Agent- and LLM-agnostic from the start. Open-source plugins cover the common agents, or roll your own: point a preset at a binary, an args template with a {prompt} placeholder, and a working directory — no plugin, no lock-in.

Watch every run live

Output streams onto the card as it happens. History keeps every run with its duration and — when the agent reports them — token and cost stats.

Run many at once

Set how many agents run in parallel (maxConcurrentAgents, 0 = unlimited). Launches over the cap wait in line and start as others finish.

Set it and walk away

Schedule runs once, daily, weekly, on an interval, or by cron — with a planner view. Keeps working in the background from the system tray.

Your machine, your keys

Everything runs locally. No account to sign up for. A managed cloud is opt-in — never required.

Also: English + French UI.

Agent presets: OpenCode and GitHub Copilot CLI with binary and args template
Define agent presets — binary, args template, working directory.
Integrations panel with a Slack webhook firing on Review, Feedback and Done
Wire webhooks that fire as cards reach Review, Feedback, or Done.
Day Planner: describe goals in plain language and generate a plan
Day Planner turns plain-language goals into a board of tasks.

How it works

Write it, wire it, schedule it — a card flows through lanes as its agent reports progress.

Draft To Do In Progress Waiting Feedback Awaiting Review Done
Write

A card carries a prompt

Add a card with a prompt, pick an agent preset and a working directory. Drafts live in their own lane until you launch them.

Myra kanban board
Connect

Bring your own agents

Claude, OpenCode, Copilot CLI, or your own — wire integrations and machines once. Myra spawns the agent headless and streams its output live onto the card.

Integrations panel
Schedule

Run on cron

Plan recurring runs from the day planner; cards advance through the lanes as agents report — no babysitting required.

Day planner

What you'll run

Real jobs people hand to Myra every day.

Nightly maintenance
Dependency upgrades

Schedule a card to bump deps, run the test suite, and open a PR while you sleep.

cron: "0 3 * * *"
On every PR
Automated review

A card watches for new pull requests and posts a structured review with findings.

agent: claude · review
Across repos
Multi-repo refactor

Fan one prompt out to several working directories and track each on its own card.

dirs: app, shared, hub

Up and running in one command

It clones every repo and wires the workspace together for you.

curl -fsSL https://raw.githubusercontent.com/Myra-Agents/Myra-Agents-Dev/develop/install.sh | bash

Prereqs: bun, Rust toolchain + Tauri v2 prerequisites. Open-source contributors get a working app without private access.

Prebuilt downloads

v0.1.0 · see all releases. Prefer to build it yourself? Use the command above.