Login or sign up
For business
Average rating
0.0/ 5

Based on 0 reviews

50%
40%
30%
20%
10%
Ask AI

Questions about kew

Based on reviews, listing data and website checks · updated 9 days ago

What is kew?

kew is an open-source control plane for autonomous coding agents that manages budgets, audit logging, approval workflows, and a kill switch outside the agent runner. It scopes work from issues to merged pull requests, dispatching tasks into isolated git worktrees with optional container-based egress controls. Users retain control over scoping and final approval decisions while the platform handles task dispatch and queue ordering by priority.

How does the six-stage workflow operate?

The process follows six stages where you scope the work, kew dispatches into isolation, works under a hard cap, a second agent reviews the first, and you keep the final judgment. Issues are ordered by priority label from p0-critical through p3-low, then by complexity which sets the step budget. A stale branch with no unique commits is auto-cleaned on re-dispatch.

How does the budget cap and circuit breaker function?

Spend is metered per issue, per day, and per week with a circuit breaker that halts dispatch when a cap is crossed. Issues estimated at or above the approval threshold are held with needs-approval instead of running. Cost is API-equivalent list price, so a local model shows $0.00 and still reports the work it did.

How does the peer review mechanism work?

A second agent attacks the first one's work by reviewing it with a fresh context that has no memory of writing the code. The reviewed commit and the merged commit are asserted equal, and findings come back as review comments. kew review <pr> dispatches an agent to address them, and the verdict is re-run against the new head.

What isolation mechanisms are used for each agent?

The platform uses git worktrees and branches named agent/issue-N for isolation, with optional container egress allowlists. Dependencies are declared by kew itself, never by the agent. A stale branch with no unique commits is auto-cleaned on re-dispatch.

How does the pricing synchronization tool operate?

Users can sync committed OpenRouter pricing snapshots or check for drift against the live API using the kew pricing sync command. The sync records the max price across provider endpoints as a conservative preflight bound, and snapshot staleness is governed by warn_age_days and max_age_days in kew.toml.

What models and concurrent operations are supported?

The system supports multiple models including Claude Code, OpenCode, pi, and local models, with a daily cap example shown as $0.00 / $40.00. It can run multiple agents simultaneously, such as three agents working on different issues like issue-1083, issue-1068, and issue-1059.

Who is the ideal user for this tool?

The platform is designed for users who want to offload coding work while keeping judgment on decisions that cost money or change the main branch. It suits teams needing budget control, audit logging, and approval workflows that operate consistently across different agents and deployment environments.

How can a user begin using the system?

To get started, users can install the CLI with pip install kew-cli and label an issue ready to enter the queue. The queue is ordered by priority and complexity, and a p0 label takes the front of the queue once an issue is marked dispatchable.

What are the stated pricing terms?

The pricing model is consumption-based and varies per provider endpoint, with OpenRouter pricing synced via the command line. The sync records per model the max price across provider endpoints, and actual per-run spend is runner-reported.

What specific integrations and technologies are named?

The tool integrates with git worktrees, branches, and optionally container-based egress controls. It also uses OpenRouter for pricing data and supports models like qwen3-coder which can be added ahead of routing.

What is the official website of kew?

The official website of kew is kew.dev.

Does kew have reviews on Trustburn?

kew has no reviews on Trustburn yet. Anyone who has used it can write the first review on its Trustburn page.

AI-generated from reviews on this page. Verify anything important with the company directly.

No reviews yet

Be the first to review kew.

Write a review

News

No news yet. This is where the business shares its own updates and press.

Photos

No photos yet. Photos shared by reviewers will show up here.

Widgets

Boost your business trust by displaying a widget on your website — a simple way to convert your visitors into customers.

Get the widget for your site →

About kew

kew is an open-source control plane for autonomous coding agents. It manages budgets, audit logging, approval workflows, and a kill switch that operate outside the agent runner, applying consistently across different agents and deployment environments. The platform scopes work from issues to merged pull requests, dispatching each agent task into an isolated git worktree and branch with optional container-based egress controls. Users retain control over scoping and final approval decisions, while kew handles task dispatch, queue ordering by priority, and step budgets.

Website
kew.dev