Orkas
A local-first desktop workforce where a Commander coordinates specialist agents for complex work.
- Source repo
- Orkas-AI/Orkas
- Stars
- ★ 9.1k
- Last updated
- today
- License
- MIT
- Primary language
- JavaScript
- FA score
- 74/100 · Some gaps
At a glance
- How it runs
- Works with
- Universal · cross-platformCodex · Claude Code · OpenAI API · Claude API
- Cost
- Free tier plus a paid hosted plan
- Setup effort
- Medium · a few setup steps
- You'll need
- Typical use
- A product manager wants to turn a PRD into actionable development tasks and distribute the work among specialists.
- Not a fit if
- Users who just want a single all-purpose chatbot
- Teams that want data hosted on a vendor cloud
- Alpine or other musl-based Linux users (unsupported)
- Source review
- 74/100 · Some gaps
What does this agent do, and when should you use it?
Orkas is an MIT-licensed, local-first desktop application for people who need coordinated agent work on complex tasks. In one chat, its Commander interprets a goal, breaks it down, and dispatches specialist agents when needed, while also handling general analysis, writing, research, file work, and automation itself. Agents can work in parallel or in sequence with focused skills, memory, and task context, while collaboration progress is shared through plan.md. The application runs on macOS, Windows, and Linux; conversations, files, knowledge bases, agent configurations, and model keys remain on the local machine, while model calls go directly to the chosen provider. It supports providers including Claude, OpenAI, Gemini, DeepSeek, Kimi, GLM, Qwen, MiniMax, and Doubao, and can drive local CLI coding agents such as Claude Code, Codex, OpenCode, and Cline.
A user submits a goal in the main chat, and the Commander uses the relevant context before staging work for specialists with dispatch_to({to, message}); recipients wake after the Commander’s current turn finishes. Every dispatch is routed through enqueue, and the Commander records collaborative progress in the shared plan.md. The main conversation is stored as jsonl, while workers see only messages matching from==me, a recipient list containing them, or mentions containing them. Each agent keeps meta/COMPETENCE.md and meta/LEARNING_STRATEGIES.md in its own directory and can use skill_manage to crystallize a solution into a private skill. Orkas can also run Claude Code, Codex, OpenCode, and Cline as local subprocesses.
- A product manager wants to turn a PRD into actionable development tasks and distribute the work among specialists.
- A researcher needs to gather, summarize, and deliver recurring market research while keeping working files on the local machine.
- A team handling sensitive local material wants to drop files into a desktop app for document chat or data analysis.
- A developer wants to coordinate Codex, Claude Code, OpenCode, or Cline from the same conversational workflow.
- A small team needs one coordinated workflow that can produce code, research, video, and slide decks using local tools.
How do you install or deploy this agent?
Packaged installers are available from orkas.ai for macOS and Windows; Linux currently requires running from source. Source execution requires Node 20+, Python 3, and macOS, Windows 10+, or recent Linux:
git clone https://github.com/Orkas-AI/Orkas.git
cd Orkas
./run.shOn Windows, use run.cmd. The scripts install dependencies and download an approximately 95 MB embedding model. After first launch, add a model-provider API key or OAuth in Settings → AI Providers.
How do you use this agent?
After first launch, configure an API key or OAuth in Settings → AI Providers, then describe the goal in the main chat. The Commander either completes general work directly or dispatches specialist agents as needed, with shared progress recorded in plan.md. Claude Code, Codex, OpenCode, and Cline can be used as CLI coding agents in the same workflow.
What are this agent's strengths and limitations?
- Local-first storage keeps conversations, files, knowledge bases, agent configuration, and keys on the machine; model calls are not proxied or archived by Orkas servers.
- It provides Commander-led multi-agent coordination through a desktop chat instead of requiring orchestration code.
- It supports multiple model providers and can mix providers across agents.
- It can incorporate Claude Code, Codex, OpenCode, and Cline as local subprocesses from the same chat workflow.
- Per-agent meta/ memory and private skills support reflection and reuse after tasks.
- Model calls still need network access unless a local model endpoint is configured; offline-capable does not make cloud models available offline.
- Users must supply their own provider API keys or OAuth and pay their model providers.
- Linux has no packaged installer and currently requires running from source.
- Source startup requires Node 20+ and Python 3, and first launch downloads an approximately 95 MB embedding model.
- The supplied documentation does not establish support for enterprise identity, centralized key management, or cloud-based team sharing.
How does this agent compare with similar options?
Compared with code-first frameworks such as LangChain and CrewAI, Orkas offers a local desktop chat interface for coordinating work. Compared with cloud agent orchestrators, it keeps data and keys on the local machine. It is distinct from a single always-on assistant such as OpenClaw or a single-assistant experience such as Claude Desktop: Orkas coordinates multiple specialists through its Commander, while OpenClaw and Hermes-Agent can serve as CLI backends.
Key facts side by side with the most closely related agents.
| Agent | Source review | Form / cost | Stars | Updated | Language | Full support on |
|---|---|---|---|---|---|---|
| Orkas This agent | 74 · Some gaps | Desktop appFreemium | ★ 9.1k | today | JavaScript | Codex · Claude Code · OpenAI API · Claude API |
| Council Decision Workspace | 82 · Good | Desktop appFree + model costs | ★ 111 | 2mo ago | Python | OpenAI API |
| NextClaw | 61 · Some gaps | Desktop appFree + model costs | ★ 264 | today | TypeScript | Codex · Claude Code |
| somora | 58 · Major gaps | CLIFree | ★ 35 | 2d ago | TypeScript | ChatGPT · Codex · Claude Code |
How does FollowAgents rate this agent?
Why each dimension lost points
The entry-point comments describe a local workspace, credential directory, and isolated skill runtime; the README says user keys are encrypted locally and excluded from sync, with model calls sent to the selected provider. It also says connectors ask before sending, posting, or deleting, and the permissions regression document specifies tests for several boundaries. Deductions: only a small code and document sample is available, so all call paths cannot be verified; rollback evidence is scattered rather than a complete recovery account; publisher identity and maintenance responsibility are only partly clear.
The README describes task budgets, checkpoints, repeated-call detection, cancellation, and catch-up runs. package.json and CI workflows list type checks and platform dependency verification, while IPC code includes error normalization and bounded diagnostics. Deductions: these files do not establish that this revision passed execution; README claims exceed the sampled source coverage, and the lockfile contents are not supplied.
The README names research, content, media, development, office, and ecommerce scenarios, and states boundaries for video recreation, Linux, offline use, packaged features, and CLI integration. Sample skills give concrete use cases and exclusions for spreadsheets, data analysis, ecommerce research, and PDFs. Deduction: trigger rules for unsampled capabilities cannot be checked for consistency.
The README is organized around capabilities, setup, architecture, FAQ, platform support, and limitations; quick start gives Node and OS requirements, commands, and native dependency repair guidance. The license is explicitly MIT. package.json provides a version, and the README marks a version threshold for automation behavior. Deductions: the packet does not provide a clear changelog or named maintenance responsibility; samples do not establish repository-wide naming consistency.
The README describes deliverable files, task delegation, project memory, and scheduled automation. Sample skills emphasize reviewable analysis, template preservation, and bounded validation stages, supporting usable outputs and value beyond single-turn chat. Deductions: the product depends on models, connectors, and first-run downloads, while built-in credits and paid tiers impose costs; no cost-benefit evidence is quantified.
The README links architecture claims to specific routing, context-budget, memory, and execution-limit source files; the permissions document defines observable test cases, and CI lists corresponding automation commands. Sample skills distinguish observed evidence from hypotheses. Deductions: many product claims remain self-reported, the supplied test sources are partial examples rather than comprehensive validation, and no tests were run to confirm behavior.
- The README describes local-first defaults, but built-in Orkas models, speech recognition, and optional cloud sync send data to services; check data flow for each feature.
- Connectors, automations, and local executable tools can cause external effects. The materials describe default confirmations and permission test plans, but this static sample cannot verify every side-effect path.
- Linux support is source-only and requires the stated glibc baseline; packaged, collaboration, and cloud features differ from the open-source source build.
Maintain this project? Add the FARS badge to your README
[](https://followagents.com/en/agents/orkas)<a href="https://followagents.com/en/agents/orkas"><img src="https://followagents.com/badge/orkas.svg" alt="FARS review"></a>Recent releases
Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.
-
v2026.11.4 Orkas 2026.11.4
Release notes excerpt
v2026.11.4 Download [Download Orkas](https://orkas.ai/download/?source=gh-orkas-release&entry_point=github_release&download=1) Highlights - Use GPT-6.1 Terra in supported provider and model settings. - Choose multiple weekdays or month dates for recurring automations and review clearer schedule summaries. - Capture product pages and recover video compositions with clearer preview and delivery checks. Fixes - Recover queued messages, selected attachments and interrupted conversations without treating downloaded files as new drafts. …View on GitHub ↗ -
v2026.10.27 Orkas 2026.10.27
Release notes excerpt
v2026.10.27 Download [Download Orkas](https://orkas.ai/download/?source=gh-orkas-release&entry_point=github_release&download=1) Highlights - Configure a complete set of public image and video models with your own provider keys, alongside existing Orkas API services. - Connect custom models through OpenAI Responses, Chat Completions or Anthropic protocols, with clearer capacity validation and credential recovery. - Create, edit and preview local apps while keeping publishing and project invitations clearly separated from local work. Improvements …View on GitHub ↗ -
v2026.10.17 Orkas 2026.10.17
Release notes excerpt
v2026.10.17 Download [Download Orkas](https://orkas.ai/download/?source=gh-orkas-release&entry_point=github_release&download=1) — Free · 500 credits · No API key needed to start. Highlights - Pick up unfinished to-dos and unread task results from a clearer Commander home with direct access to connectors. - Manage local projects with clearer action menus, reliable navigation, and more consistent task boards. - Clean up older tasks with age filters, explicit selection, running-task confirmation, and reversible recovery. Improvements …View on GitHub ↗
FAQ
Is Orkas a local LLM?
What does it cost?
Does Orkas send my data to its servers?
Can it work fully offline?
Thanks — we use this to improve the page.