Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.mirage.strukto.ai/llms.txt

Use this file to discover all available pages before exploring further.

Mirage architectureMirage architecture
Mirage stacks four thin layers between an agent and the services it touches.

1. AI Agent and Application

The agent (or any application embedding Mirage) issues bash commands, VFS calls, or syscalls. One vocabulary, every backend.

2. Mirage Bash and VFS

The action surface. Mirage Bash parses commands with tree-sitter and runs them against the Mirage VFS, a unified filesystem API over every mount. A FUSE Adapter exposes the same tree to host tools when you want it. A Command Registry and VFS Registry describe what verbs and resources are available.

3. Dispatcher & Cache

The Mirage Dispatcher routes each operation to the mount that owns the path, joining pipelines that span systems. The Index & File Cache absorbs repeats: the first directory walk hits the API, the next serves from cache; the first read streams bytes, later reads are local.

4. Infrastructure and Remote

Whatever you mount: RAM, Disk, Redis, S3 / R2 / GCS / OCI / Supabase, Gmail / GDrive / GDocs / GSheets / GSlides, GitHub / Linear / Notion / Trello, Slack / Discord / Telegram / Email, MongoDB, SSH, and more. Each speaks the same filesystem semantics from the agent’s point of view. Browse the Resource Matrix for the full list.

Where to go next