# Reversi

- [Reversi](https://reversi.mintlify.app/index.md): Reversibility, approvals, policy and tamper-evident evidence for every tool an AI agent can touch.
- [Quickstart](https://reversi.mintlify.app/quickstart.md): Journal an agent's tool calls, then roll them back, in about five minutes.
- [Manual testing guide](https://reversi.mintlify.app/testing.md): Step by step, copy-paste instructions to test every feature of Reversi by hand, with the output to expect.
- [Manual testing - adoption](https://reversi.mintlify.app/testing/adoption.md): Try by hand, in one terminal, everything that takes a team from "an MCP server we do not trust yet" to enforced, signed manifests kept in git.
- [Manual testing - policy, approvals and evidence analysis](https://reversi.mintlify.app/testing/governance.md): Hand-test rich manifests, policy rules, budgets, freezes, approvals, delegation, halts, simulation, replay and the console editors, with the output to expect.
- [Operations, security and integrations](https://reversi.mintlify.app/testing/operations.md): Hand-test the console modes, API keys, OIDC, notifications, alerts, tracing, evidence, retention, egress rules and the three stores, with the output to expect.
- [Tool manifests](https://reversi.mintlify.app/concepts/manifests.md): How a tool declares its reversibility, its undo, who owns it and who must approve it, plus the registry, lint and contract tests that keep manifests honest.
- [Rollback](https://reversi.mintlify.app/concepts/rollback.md): Saga rollback with dependency ordering, retries, verification, residue reporting and honest reporting.
- [Ledger](https://reversi.mintlify.app/concepts/ledger.md): The tamper-evident record of everything that happened.
- [Policy as code](https://reversi.mintlify.app/concepts/policy.md): Rules, limits, budgets, freezes, time windows, risk scoring, approvers, versions and tests that add restrictions on top of manifests and scopes.
- [Approvals](https://reversi.mintlify.app/concepts/approvals.md): Human decisions bound to the exact call, used once, expiring, routed to the right people, and recorded.
- [Simulation and capture](https://reversi.mintlify.app/concepts/simulation.md): Try a policy on last week's calls, replay a run with one argument changed, preview what approving a call does, and compare plans before running any of them.
- [Content checks](https://reversi.mintlify.app/concepts/content-checks.md): Heuristics for credentials, hidden payloads, prompt-injection phrasing and data exfiltration.
- [Evidence](https://reversi.mintlify.app/concepts/evidence.md): Signed bundles, anchored ledger heads, run and period reports, and independent verification.
- [Security model](https://reversi.mintlify.app/concepts/security-model.md): Roles, approvals, scopes and what Reversi assumes.
- [MCP gateway](https://reversi.mintlify.app/gateway.md): Journal, police, approve and roll back any MCP server's tool calls, without changing the agent or the server.
- [Connectors](https://reversi.mintlify.app/connectors.md): Ready-made tools for the filesystem, GitHub, GitLab, Slack, Teams, Linear, Jira, Stripe, Google Calendar, Postgres, S3, Cloud Storage, Blob Storage and Kubernetes, each with a Reversi manifest, a compensation that really undoes it, and a certification suite.
- [Framework integrations](https://reversi.mintlify.app/integrations.md): Put LangGraph, CrewAI, OpenAI Agents SDK, FastMCP and Claude Code tool calls behind Reversi with the Python SDK adapters.
- [Operator console](https://reversi.mintlify.app/console.md): See what an agent changed, decide approvals, run rollbacks, edit manifests and policy, and download evidence in a browser.
- [CLI](https://reversi.mintlify.app/cli.md): Set up a project, lint and test manifests, publish and apply them, inspect and roll back runs, decide approvals, simulate policy, sign evidence.
- [TypeScript SDK](https://reversi.mintlify.app/sdks/typescript.md): @multimindlab/reversi: the engine, embedded in your process.
- [Python SDK](https://reversi.mintlify.app/sdks/python.md): reversi-sdk, a client for the Reversi server, offline ledger and manifest checks, agent-framework adapters. Standard library only.
- [Trace viewer](https://reversi.mintlify.app/viewer.md): A single HTML file that verifies a ledger and shows the timeline.
- [Deployment](https://reversi.mintlify.app/operations/deployment.md): Run the server and gateway in Docker, Docker Compose or Kubernetes (Helm), configure them, and put TLS in front.
- [GitOps](https://reversi.mintlify.app/operations/gitops.md): Keep manifests and policy in git, check them in CI, and make a registry match with `reversi apply`. The project file, hot reload, and why there is no Terraform provider.
- [Storage](https://reversi.mintlify.app/operations/storage.md): JSON files, SQLite or PostgreSQL for the ledger, run state and approvals.
- [Authentication and roles](https://reversi.mintlify.app/operations/authentication.md): API keys with rotation and groups, OIDC bearer tokens, console modes, and what each role may do.
- [Monitoring](https://reversi.mintlify.app/operations/monitoring.md): Health, Prometheus metrics, structured logs, alerts and OpenTelemetry traces.
- [Notifications](https://reversi.mintlify.app/operations/notifications.md): Tell people about approvals, failed rollbacks and alerts through Slack, Teams, email, PagerDuty, Opsgenie or a signed webhook.
- [Retention, holds and data residency](https://reversi.mintlify.app/operations/retention.md): How long run data is kept, how a legal hold stops deletion, where Reversi may send data, and how personal data is masked.
- [API overview](https://reversi.mintlify.app/api-reference/overview.md): HTTP API conventions, roles, status codes and an index of every route.
- [Agent endpoints](https://reversi.mintlify.app/api-reference/runs.md): Open runs, request calls, report outcomes. Agent role.
- [Operator endpoints](https://reversi.mintlify.app/api-reference/operator.md): Plan, roll back, resolve, export, report, search, preview and simulate. Auditor reads; operator acts.
- [Approvals, evidence and admin endpoints](https://reversi.mintlify.app/api-reference/admin.md): Decide approvals, download signed bundles, flip kill switches, manage manifests, policy, keys, holds and retention.
- [Threat model](https://reversi.mintlify.app/threat-model.md): What Reversi defends against, what it assumes, and what it does not cover.
- [Limits](https://reversi.mintlify.app/limits.md): What Reversi does not do, and what is built but not yet checked against real services. Read this before relying on it.
- [Roadmap](https://reversi.mintlify.app/roadmap.md): What is built, what is built but unverified against real systems, and what is not built.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.