Archal

Archal provides stateful API sandboxes for testing AI agents and integrations. Each sandbox holds one or more simulated environments that behave like the services an agent calls in production, keeping state between calls and resetting to a declared baseline, so tests run without shared test accounts, rate limits or leftover data. It is built for agent developers, connector teams and integration engineers. Environments cover GitHub, GitLab, Linear, Jira, Slack, Discord, Stripe, Supabase, HubSpot, Datadog, Sentry, Google Workspace, ClickUp, Cal.com, Customer.io, Ramp, Webflow and more, most available over both REST and MCP. The workflow is create, call, inspect and reset: a sandbox starts from sample state, custom state or a blank environment, the agent uses familiar API paths and tools, and requests, responses, mutations and timing can be inspected together. Typical uses are harness testing, giving each branch or CI job its own sandbox, and reinforcement learning rollouts that need an identical starting state. A setup skill can be pasted into Claude Code, Codex, Cursor or Devin to handle signup and configuration. Pricing is usage based with no subscription and no card needed to start. Each environment costs $0.10 per minute, prorated by the second. New accounts receive $20 in usage credits: $5 after verified signup and $15 more when the first sandbox is ready. Each environment is an independently built simulation of the service it represents, not the provider's live service, so no test traffic reaches the real provider and coverage follows the documented operations.