StateWeave | Test AI agents on realistic copies of your apps
New
Sandboxes for AI agents
Your agent’s apps, cloned .
Deploy sandboxes with stateful clones of Slack, Salesforce, Zendesk and more: the same APIs, MCP servers and UIs your agents use. Test on a replica of your full system before production.
Get started for free
Book a demo
Get started with your AI
Claude Code
Codex
Cursor
+ Your agent uses an app we don't have? We'll clone it for you
Ridgeline Brew Works
· 4 twins
agent@0.3 running
Example
agent
users.messages.send
200
How it works
From your agent to a graded run.
Six steps, most of them one command. Click a step, or use the arrow keys.
Twins
Every app your agent touches.
Each twin has the same API, events, MCP server and UI as the real app. Don't see yours? We'll build it.
Slack
Salesforce
Gmail
Zendesk
Shopify
Google Sheets
Google Calendar
Google Drive
Teams
Stripe
HubSpot
GitHub
Ready today
Need an app that isn't here? We'll clone it.
Tell us which app your agent works in, and we build its twin for you.
1 You name the app and what your agent does in it
2 We build its twin with the same API, events and web UI
3 We check it against the real app, in a test account
It's in your workspace ready for sandboxes and tasks
T e s t t h e n e x t c h a n g e b e f o r e i t s h i p s
Discuss your use case and see how StateWeave can support your team.
Get started for free
Book a demo
Or email us at support@stateweave.net
A walkthrough of the twins your agent needs
A scenario built around your use case
How to test every model, prompt and harness change
Get started with your AI Claude Code Codex Cursor
01 Sign in
01 · Sign in
Make a key in your dashboard.
In your StateWeave dashboard, open Connect your agent and click Create key. It gives you the commands to paste, with the key already in them.
Quickstart →
02 Connect
02 · Connect
Your coding agent connects your agent.
sw setup gives Claude Code, Codex or Cursor StateWeave's skills. Ask it to set StateWeave up: it finds the apps your agent uses and points them at fresh copies. Your agent keeps its own SDKs.
From your coding agent →
03 Sandbox
03 · Sandbox
Fill a sandbox with your company.
Write to the apps through their own APIs, or describe your company in a sentence and StateWeave generates it. Keep it as a version: every run starts from a fresh copy.
Sandboxes →
04 Tasks
04 · Tasks
Write tasks, with the failures you fear.
A task is a folder: what your agent is asked, and checks that read the apps through their own APIs. Add faults like a Slack outage. sw task import proves a task before keeping it.
Tasks →
05 Run
05 · Run
Run your agent on fresh copies.
sw run hands your agent fresh copies of the sandbox's apps, with the faults the task asks for. Your agent runs where it always does, and every call it makes is recorded.
Run your agent →
06 Results
06 · Results
Graded on what changed.
After the run, the checks read the apps through their own APIs. See which failed and every call behind it, fix your agent, and run the same task again.
Test runs →
Additional pricing (as published)
Orders
| Order | Date | Customer | Total | Payment status | Fulfillment status |
|---|---|---|---|---|---|
| #1431 | Oct 14 at 9:12 am | Olivia Park | $289.00 | Paid | Refunded |
| #1428 | Oct 14 at 8:40 am | Ben Wong | $289.00 | Refunded | Fulfilled |
| #1419 | Oct 13 at 6:05 pm | Hassan Patel | $318.50 | Paid | Fulfilled |
| #1402 | Oct 12 at 2:31 pm | Emily Young | $289.00 | Paid | Fulfilled |