# Doppels

## Docs

- [Doppels: Freeze AI Agent Sessions into Replayable Recipes](https://docs.doppels.so/introduction.md): Doppels converts AI agent sessions into typed, validated, replayable local Capabilities — same inputs, same steps, same auditable result every time.
- [Install the Doppels CLI on macOS, Linux, or from Source](https://docs.doppels.so/installation.md): Install the Doppels CLI on macOS or Linux using the curl installer, Homebrew tap, or by building from source with mise. The binary is named `doppels`.
- [Doppels Quickstart: Validate, Run, and Inspect a Capability](https://docs.doppels.so/quickstart.md): Run your first Doppels Capability in minutes: initialize a Space, validate manifests, execute a run, and inspect logs — no AI agent required.
- [How Doppels Works: Freeze Once, Replay Forever](https://docs.doppels.so/concepts/how-it-works.md): Understand the Doppels execution model: agent exploration, the freeze step, YAML manifests committed to Git, and zero-token local replay.
- [Capabilities: The Public Contract for Your Automation](https://docs.doppels.so/concepts/capabilities.md): A Capability declares what an automation does — its inputs and outputs — as a versioned YAML manifest you commit to your repository.
- [Recipes: The Local Execution Plan for a Capability](https://docs.doppels.so/concepts/recipes.md): A Recipe implements a Capability — declaring shell steps, required tools, output mappings, and approval rules that execute locally on your machine.
- [Spaces: Scoping Your Capabilities and Recipes](https://docs.doppels.so/concepts/spaces.md): A Space is the configuration boundary for a Doppels project — scoping which Capabilities and Recipes are discovered and how runs are stored.
- [Freeze an AI Agent Session into a Doppels Capability](https://docs.doppels.so/guides/freeze-with-agent.md): Use the doppel-freeze skill in Claude, Cursor, or Codex to capture a working agent session as a replayable YAML Capability and Recipe.
- [Validate and Run Capabilities with the Doppels CLI](https://docs.doppels.so/guides/validate-and-run.md): Use doppels validate to check your manifests and doppels run to execute a Capability locally with full audit trail under .doppels/runs/.
- [Share Doppels Capability Run Results with Your Team](https://docs.doppels.so/guides/sharing.md): Use Doppels Share to expose a Capability result to collaborators — only declared outputs are shared, never your credentials or local scripts.
- [Using Git to Version and Share Doppels Capabilities](https://docs.doppels.so/guides/git-workflow.md): Commit your Capability and Recipe YAML files to Git to version, review, and share automations across your team like any other source code.
- [Doppels Agent Skills: Teach Your AI to Freeze Sessions](https://docs.doppels.so/skills/overview.md): Agent skills extend Claude, Cursor, and Codex with Doppels capabilities — install them once and trigger freeze from any agent conversation.
- [doppel-freeze: Freeze Agent Sessions into YAML Recipes](https://docs.doppels.so/skills/doppel-freeze.md): The doppel-freeze skill teaches your agent to capture any working session into a versioned, typed Capability and Recipe you can replay locally forever.
- [Doppels CLI Reference: All Commands and Flags Explained](https://docs.doppels.so/reference/cli.md): Complete reference for every doppels CLI command — spaces init, validate, capabilities list, describe, run, runs show, logs, and telemetry with all flags.
- [Doppels YAML Schema Reference: Capability, Recipe, Space](https://docs.doppels.so/reference/yaml-schemas.md): Complete YAML schema reference for Capability, Recipe, and Space manifests — the three document types that define automations in Doppels.
- [Doppels Template Expressions: Wiring Inputs and Outputs](https://docs.doppels.so/reference/expressions.md): Use {{ inputs.name }} and {{ steps.id.result }} expressions in env, produces paths, and returns fields to wire inputs and outputs between steps.
- [Doppels FAQ: Execution, Tokens, Secrets, and Licensing](https://docs.doppels.so/reference/faq.md): Answers to common questions about Doppels: execution model, tokens, secrets, Capability vs Recipe, offline use, and open-source licensing.