devvy
Log inGet started

Background agents
to meet your flow

Deploy AI coding agents that work in the background. responding to comments, shipping PRs, and running tests without interrupting your team.

No credit card·BYOK model·Unlimited agents
Devvy
App
Today
Message Devvy
Setup in minutes

Configure once.
Run everywhere.

Four steps to your first autonomous PR. No infrastructure to manage. Agents run in isolated VMs on our compute.

Step 01

Choose your runtime

Pick the AI coding agent that powers your background workers. Each has different strengths; mix and match per task type.

Claude CodeClaude Code
OpenCodeOpenCode
OpenAI CodexCodex
GeminiGemini CLI

Intelligent guardrails,
not just automation.

Agents that blindly ship create risk. Devvy gives you precise control over when humans stay in the loop, and when agents run free.

devvy.yaml
1runtime: claude-code
 
3# human approval checkpoints
4require_approval: true
5approval_phase: post-plan
6approval_timeout: 24h
 
8# automated test gates
9test_gates:
10 - run: npm test
11 on_fail: retry_once
12 - run: npm run lint
13 on_fail: abort
 
15# per-repo configuration
16token_budget: 200000
17escalate_on: timeout
18excluded_paths:
19 - node_modules/
20 - dist/
21model_overrides:
22 simple: gpt-4o-mini
23 complex: claude-opus-4

Humans stay in the loop

Devvy pauses after the planning phase and posts a plain-English summary of what it intends to do. You approve, modify, or abort before a single line of code is written.

Configurable pause point: post-plan, pre-execute, or both
24-hour approval timeout with Slack notification
Full plan diff visible in your chat thread
Bring Your Own Keys

You own your AI costs.
We just run the compute.

Connect your own API keys for Claude, GPT-4, Gemini, or any Bedrock model. Token costs go directly to your provider accounts. Devvy never marks up LLM usage.

No LLM markup

Pay your provider's published rates directly. Devvy never adds margin on tokens. Costs go straight to your Anthropic, OpenAI, Google, or AWS account.

Switch models per task

Use fast, cheap models for simple fixes and powerful ones for complex refactors. Model selection is a one-line change in devvy.yaml.

Full cost visibility

Per-run token and compute costs are visible in your dashboard. No surprise bills. See exactly what each agent task cost before and after.

Keys encrypted at rest

AES-256 encryption, never written to logs or audit trails, never included in VM snapshots. Rotate or revoke any key instantly from the dashboard.

Review response

Responds to review
comments instantly.

When a reviewer requests changes, Devvy reads the comment, understands the context, writes the fix, and pushes a new commit, all without interrupting your engineer.

Reads every review comment
Understands the surrounding code context
Writes the fix in an isolated VM
Pushes a new commit and resolves the thread
PR #142 · fix: add null check for userId in getUseropen
ACTIVE AGENTS
0
NO LIMITS
Fix: null check in auth middlewarefix/auth-null
Add pagination to /api/listingsbackend
Write unit tests for CartServicefrontend
Migrate user table to new schemainfra
Refactor analytics event pipelinebackend
Update Stripe webhook handlerpayments
Concurrency

Run every ticket
concurrently.

Each agent runs in its own isolated VM. Spin up one or fifty; there are no queues, no concurrency limits, no extra configuration needed. Your backlog clears itself.

∞
Concurrent agents
<2m
Median task time
0
Config required
Each VM boots in under 2 seconds
Agents never share state or file systems
VMs are destroyed immediately after the task
Pricing

Simple, transparent
pricing.

Free

For individuals exploring Devvy.

$0
Free forever
Get started free
  • Up to 2 engineers
  • 25 agent runs / month
  • GitHub integration
  • Slack & Discord triggers
  • Community support
Most popular
Pro

For teams that ship every day.

$5/ mo
Start 14-day trial
  • Unlimited engineers
  • Unlimited agent runs
  • GitHub, GitLab & ADO
  • Slack, Discord, Teams, WhatsApp
  • YAML workflow config
  • Human approval checkpoints
  • Priority support
Enterprise

Custom contracts and dedicated SLAs.

Custom
Talk to sales
  • Everything in Pro
  • SAML / SCIM / SSO
  • Dedicated VPC compute
  • On-prem deployment option
  • Custom SLA
  • Dedicated success manager

All plans include BYOK. You only pay for compute, never for LLM tokens.

FAQ

Common questions.

Still have questions? Email us

Let Devvy
write the code.
You ship.

No credit card required. Free plan available.

182ms
avg VM boot time
100%
isolated environments
0
context switches