{"slug": "worca-deterministic-multi-agent-pipeline", "title": "Worca – Deterministic Multi-Agent Pipeline", "summary": "Worca launched a deterministic multi-agent pipeline that runs coding agents through Clarify, Plan, Refine, Implement and Review stages, looping until the work passes its quality gates and ending on its own git branch one click from a pull request. Worca installs via `npm i -g @worca/app` on macOS, Linux or native Windows, runs agents inside a disposable Linux container without the home folder mounted, and supports Claude, GitHub Copilot, Ollama, LM Studio and free OpenRouter models with per-run caps and budgets. The tool records every step, verdict and file in a single SQLite database in the user's home directory, keeps run state out of the repo, and offers a mock mode that runs the whole pipeline offline with no agents or tokens for demos, development and CI.", "body_md": "# Your AI agents,\n\non rails.\n\nDescribe the work, get back a **reviewed pull request**. A **deterministic engine** runs the agents, loops until the work passes its quality gates, and stops only to ask **you** what matters.\n\n## Describe the work.\n\nApprove the result.\n\nAgents do the creative work; the engine does the control flow. Pick a workflow, or let **Auto** build one for the task. Every step, verdict and file is recorded.\n\n**Clarify**\n\nHidden decisions become questions before any code is written.\n\n**Plan**\n\nThe planner explores the codebase and writes a concrete plan.\n\n**Refine**\n\nThe refiner rewrites the plan until no major issues remain.\n\n**Implement**\n\nThe implementer follows the approved plan, test first.\n\n**Review**\n\nThe reviewer sends blocking findings back until the diff is clean.\n\nA finished run ends on its own branch, **one click from a pull request**. [See the whole pipeline →](https://worca.dev/how-it-works/)\n\n## Everything\n\na run needs.\n\nAny model, schedules, a record of every run, limits you set outside the agent, and a whole team working in one place.\n\n[How it works](https://worca.dev/how-it-works/)\n\n**Workflows you can see and change.**\n\nAuto picks the workflow, or build one in the Composer: agents, script cards and bounded loops. Every step files what it wrote.\n\nThe pipeline →\n[Models & costs](https://worca.dev/models/)\n\n**Any model. Every dollar accounted.**\n\nClaude, GitHub Copilot, Ollama or LM Studio on your desk, free OpenRouter models. Caps, budgets, and the hours each run saved.\n\nModels & costs →\n[Automation](https://worca.dev/automation/)\n\n**Runs that start while you sleep.**\n\nSchedule a run for tonight or every weekday, chain runs onto each other's branches, and answer from Slack or Telegram.\n\nAutomation →\n[Ask Worca](https://worca.dev/ask-worca/)\n\n**An assistant that knows your runs.**\n\nAsk what the reviewer flagged, attach a screenshot, let it read the docs page, or have it schedule the next run.\n\nAsk Worca →\n[Teams](https://worca.dev/teams/)\n\n**One Worca for the whole team.**\n\nMetrics and a delivery timeline, a team policy kept in git, a shared server behind sign-in, and each person pays with their own key.\n\nTeams →\n[Security](https://worca.dev/security/)\n\n**Autonomy, with limits.**\n\nGuardrail policies that repo settings can't undo, a disposable container, and agents that never see a model key.\n\nSecurity →\n## From your laptop\n\nto your team's server.\n\nThe same Worca everywhere. Start on your own machine and move to a shared server when the team joins, keeping your workflows and history.\n\n**One npm install**\n\nmacOS, Linux or native Windows. `npm i -g @worca/app`\n\n**A disposable Linux box**\n\nAgents run inside it, and your home folder isn't mounted. `worca container up`\n\n**One Worca, behind sign-in**\n\nAny container host, behind the sign-in you already use. Worca records who started, paused and approved each run.\n\n**Metrics and policy in git**\n\nSpend, a delivery timeline and team caps, stored on branches in your own repo. No extra server.\n\n## The rest of\n\nthe toolbox.\n\n**Isolated worktrees** Each run gets its own git worktree and branch. Your checkout is never touched; parallel runs don't collide.\n\n**Runs pause, they don't die** A 500, a rate limit or an expired login pauses the run with its worktree intact. Resume picks up where it stopped.\n\n**Mock mode** The whole pipeline runs offline with a deterministic mock: no agents, no tokens. For demos, development and CI.\n\n**Plugins & marketplaces** Task sources, agents, scripts, skills, workflows, models and chat channels. Each install shows what it adds and asks you first.\n\n**Workspaces** Group related repos. One run fans out across all of them with a single cross-project review verdict.\n\n**One-click PRs** A finished run shows its diff and opens a pull request, even from a fork, and reports back to the issue it came from.\n\n**Nothing in your repo** Run state lives in one SQLite database in your home directory. Your checkout stays clean.\n\n**Simple to Expert** New installs start with a short sidebar and a Getting started checklist. Switch to Advanced or Expert when you want everything.\n\n## Ship work you\n\ndidn't babysit.\n\nOne npm install on macOS, Linux or Windows. Or a container, or a shared server for the team.", "url": "https://wpnews.pro/news/worca-deterministic-multi-agent-pipeline", "canonical_source": "https://worca.dev/", "published_at": "2026-10-24 07:57:48+00:00", "updated_at": "2026-10-24 08:12:25.143298+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "developer-tools", "ai-products"], "entities": ["Worca", "Claude", "GitHub Copilot", "Ollama", "LM Studio", "OpenRouter", "Slack", "Telegram"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/worca-deterministic-multi-agent-pipeline", "markdown": "https://wpnews.pro/news/worca-deterministic-multi-agent-pipeline.md", "text": "https://wpnews.pro/news/worca-deterministic-multi-agent-pipeline.txt", "jsonld": "https://wpnews.pro/news/worca-deterministic-multi-agent-pipeline.jsonld"}}