# strukto-ai/mirage

The World's First Unified Virtual Filesystem For AI Agents

Repository: https://github.com/strukto-ai/mirage
Canonical: https://ross.abutalabs.com/products/mirage
Homepage: https://www.strukto.ai/mirage
Language: TypeScript
License: Apache-2.0
License Family: permissive
Topics: agent-sandbox, agent-tools, ai-agents, bash, claude-code, fuse, llm-agents, openai-agents, python, typescript, vfs, virtual-filesystem, dsh-plugin, dsh
Last push: 2026-08-26T22:12:45+00:00

## Health v2 (maintenance only)
Score: 80/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 98, longevity 8
- inputs: {"age_days": 119, "days_push": 7, "days_rel": 19, "gap_med": 22.0, "n_releases_24m": 5}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3569, forks 260 (observed 2026-08-28T04:08:10.058643+00:00)

## What it is
Mirage is a unified virtual filesystem for AI agents that mounts services like S3, Google Drive, Slack, Gmail, Postgres, and Redis side-by-side as one filesystem. Agents use standard bash commands (cat, grep, pipes) to read, write, and pipe data across all backends, with snapshot/rollback versioning and a two-layer cache.

## Use cases
- give llm agents file access to s3 and slack via bash
- mount cloud services as a virtual filesystem for ai agents
- let claude code or openai agents grep across s3 and google drive
- sandbox agent file operations with snapshots and rollback
- pipe data between postgres, redis, and slack in one shell command
- embed a virtual workspace in fastapi or express for agent runs
- cache repeated agent reads from cloud backends

## When to choose
- you're building LLM agents that need uniform file-style access to multiple SaaS and cloud backends
- you want agents to reuse bash skills instead of learning per-service APIs
- you need sandboxed, versioned, portable agent workspaces

## When to avoid
- you need a general-purpose production filesystem or FUSE mount for human users
- your stack is outside Python/TypeScript/Node ecosystems
- you only need a single backend integration with no agent context

## Facets
- artifact type: library
- maturity: active
- function: file-system, cli, agent-framework, caching, sdk, developer-tools
- domain: large-language-models, developer-tools, files
- platform: python, cross-platform, cli, browser
- tags: virtual-filesystem, vfs, bash, agent-sandbox, fuse, workspace-snapshots, s3, slack, google-drive, redis, llm-tools, ai-agents, automation, nodejs, typescript

## Member repositories
- strukto-ai/mirage (main) score 80

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:10.058643+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:34:08.699815+00:00, confidence not recorded.
  - readme: https://github.com/strukto-ai/mirage (fetched 2026-08-28T04:08:10.058643+00:00, sha 14969ebb615a)
  - homepage: https://www.strukto.ai/mirage (fetched 2026-08-29T09:28:03.448113+00:00, sha 3c4d77cd362b)
  - site_page: https://www.strukto.ai/about (fetched 2026-08-29T09:28:03.457366+00:00, sha ffb6c54adf0d)
  - site_page: https://docs.mirage.strukto.ai (fetched 2026-08-29T09:28:03.459526+00:00, sha 8b19b92206f2)
- Data as of 2026-08-30T08:39:29.467469+00:00.