Give your AI agent a goal. It plans the work, builds it, checks itself, challenges the result, and prepares a safe delivery. You get a visual workflow and a clear history instead of a wall of terminal output.
Get started · See what it does · Open Workflow Studio · Read the docs
AI coding tools are fast, but important details can disappear between the prompt, the code, and the final answer. AI Engineering Loop gives every task one calm, repeatable path:
| You want | The Loop handles |
|---|---|
| A clear outcome | Turns the request into an agreed Goal before code changes |
| Less supervision | Guides your agent through planning, building, and checking |
| Honest progress | Shows what is ready, running, blocked, or waiting for approval |
| Fewer surprises | Tests happy paths, edge cases, and failure cases |
| A trustworthy result | Compares every claim with real evidence before delivery |
| A useful history | Keeps searchable Runs, decisions, artifacts, and proof |
You stay in control of the product decisions. Your AI agent takes care of the repetitive engineering ceremony.
Install it in any project:
npx ai-engineering-loop initThen tell your coding agent:
Use AI Engineering Loop to build this feature: your idea here
That is the intended experience. You do not need to memorize the internal commands—the agent can operate the loop for you.
Want the visual workspace?
npx ai-engineering-loop studioWorkflow Studio opens locally in your browser. Your project and run evidence stay on your machine.
On first setup, .ai-engineering-loop/ is added to your project’s .gitignore so plans and Run evidence stay private by default. Want to share that context with your team? Simply remove that line from .gitignore.
Workflow Studio makes agent work feel like a product, not a terminal session.
- Connect steps on an open canvas
- Add the next step directly from any node
- Rename nodes in place without breaking their identity
- Select, move, duplicate, or delete groups of nodes
- Turn selected steps into independent Loop Groups
- Label repeat and continue paths in plain language
- Import and export workflow JSON safely
- Create your own prompt-powered Agent nodes
Write the task in plain language, add the outcomes that matter, and Freeze the Goal when it is ready. The agent cannot quietly move the finish line after work begins.
Runs have human-friendly names instead of cryptic hashes. Search previous work, open any Run as a read-only workflow, and see exactly where it stopped.
Select a node to inspect its real Input, Output, Evidence, and timeline. The canvas shows the actual status of every step, so completed work, failures, approvals, and waiting nodes are easy to understand.
Use Copy for AI Agent to create a ready-to-use handoff for the selected Run. The agent continues with the same Goal, workflow, and evidence instead of starting from scratch.
If the agent needs clarification, its questions can appear inside Workflow Studio. You answer them there, and the answer stays attached to the Run.
Create multiple independent Loop Groups in one workflow:
- Select the steps that belong together
- Choose where the loop starts and where the decision happens
- Give repeat and continue paths clear labels such as
NO · RETRYandYES · CONTINUE - Set a maximum number of passes
- Follow a separate
Iteration 2 / 3counter for each group
When a loop reaches its limit, Workflow Studio pauses for approval. You can allow one focused extra pass, continue with the current evidence, or stop the Run safely. Nested and overlapping loops stay disabled so the workflow remains easy to read.
The Execute button lives beside the workflow trigger. A Run starts only when the Goal, workflow, mode, and safety checks are ready.
New in v1.9.0: Run Checkout, AI Agent handoff, live Run questions, and Multiple Loop Groups. See the release.
The experience is organized into four simple moments:
- Specify — agree on what success and failure look like.
- Make — let the agent investigate, plan, and build.
- Review — verify the result and challenge unsupported claims.
- Deliver — ask for human approval before publishing.
Under the hood, these moments use eight evidence-backed stages. The rigor is there when you need it, but it does not have to dominate the experience.
Specify (stages 0-1) → Make (stages 2-4) → Review (stages 5-7) → Deliver (stage 8)
AI Engineering Loop is designed to work with the tools developers already use:
- Claude Code
- Grok CLI
- Gemini / Antigravity
- Other agent hosts that can read project instructions and run local commands
The project includes ready-to-use skills and reviewer roles for supported hosts. Each host follows the same Goal, verification, review, and approval rules.
Start from a built-in workflow or make your own:
- Product feature
- Bug fix
- Security review
- Repository audit
- Refactor
- Custom prompt-powered Agent workflow
Custom Agents can reference approved capabilities, MCP connections, input artifacts, and output formats. They do not receive hidden shell, filesystem, network, or secret access.
The Loop keeps a few promises:
- Nothing ships only because an AI says “done.”
- Goals are frozen before implementation.
- Verification records the real command and result.
- A Devil’s Advocate challenges the work before the Judge decides.
- Sensitive context is bounded and redacted.
- Publishing in assisted mode needs human approval.
- The browser never silently runs arbitrary shell commands.
Choose the level of autonomy that feels right:
| Mode | Best for |
|---|---|
| Report only | Explore a codebase without changing it |
| Assisted | Everyday work with human approval before delivery |
| Unattended | Trusted automation with stricter isolation and limits |
Assisted mode is the default. Unattended mode stays off until you deliberately enable it.
The friendly workflow sits on top of a complete local CLI. You can inspect recipes, budgets, evidence, nodes, gates, policies, and handoffs whenever you need to.
Show common CLI commands
# Project setup and health
npx ai-engineering-loop init
npx ai-engineering-loop status
npx ai-engineering-loop doctor
# Start work
npx ai-engineering-loop run "describe the task"
npx ai-engineering-loop studio
# Explore workflows
npx ai-engineering-loop recipe list
npx ai-engineering-loop recipe simulate default --mode assisted
# Inspect the active Run
npx ai-engineering-loop state --json
npx ai-engineering-loop node status
# Emergency stop and resume
npx ai-engineering-loop budget pause
npx ai-engineering-loop budget resumeSee the core/ documentation for the full runtime, evidence, recipe, security, and adapter contracts.
Your living project context and Run evidence are stored under .ai-engineering-loop/ and kept out of normal commits by default. You can remove its .gitignore entry whenever you want to share that context. Delivery adapters support:
- Standard Git
- GitHub
- GitLab
- Custom team workflows
The same engineering flow stays consistent even when teams use different delivery tools.
The rules that decide whether AI-generated work is trustworthy should be inspectable. AI Engineering Loop keeps its schemas, safety policies, verification logic, recipes, and agent prompts in the repository.
No hidden cloud control plane is required. No project source is uploaded by Workflow Studio.
- Workflow Studio
- Workflow recipes
- Runtime safety
- Verification
- Controlled workflow runtime
- Security
- Support
Ideas, workflow recipes, bug reports, and pull requests are welcome. Start with an issue or open a focused PR with a clear explanation of the user outcome.
npm test
npm run doctorMIT © Ega Gofur
Let the agent handle the process. Keep your attention on the product.