Skip to main content

FrootAI — AmpliFAI your AI Ecosystem Get Started

← Back to the Orchard

FAI Orchard — Getting Started

FAI Orchard — Getting Started

Welcome. This is the 5-minute quickstart for FAI Orchard v1.0 — the MCP-native catalog of cross-cloud AI accelerators that ships byte-equal across npm, PyPI, and a signed self-hosted Docker image.

ℹ️ TL;DR: pick npm OR PyPI (both give byte-equal results), drop one JSON snippet into your MCP client config, and your agent can run `orchard.search` + `orchard.show` against the 413-entry catalog.

§1 — Install in your MCP client

Pick the runtime you prefer. Both produce byte-equal results across the 27 contract assertions in our cross-runtime conformance suite.

Option A — npm (Node ≥ 18)

```bash npx -y @frootai/[email protected] ```

Option B — PyPI (Python ≥ 3.10)

```bash uvx frootai-mcp-orchard[sdk]@1.0.0 ```

Option C — Self-hosted Docker

```bash docker pull ghcr.io/frootai/mcp-orchard:1.0.0 docker run --rm -p 3000:3000 ghcr.io/frootai/mcp-orchard:1.0.0 ```

The Docker image is multi-stage Alpine, runs as non-root (UID 1001), ships with HEALTHCHECK + cosign signatures + CycloneDX SBOM attestation. See Self-host guide for the full deploy playbook.

§2 — Wire into your MCP client

Drop this JSON snippet into your client config:

```json { "mcpServers": { "frootai-orchard": { "command": "npx", "args": ["-y", "@frootai/[email protected]"] } } } ```

For per-client install snippets (Cursor / Claude Desktop / ChatGPT MCP / Continue / Cline / VS Code), see Install in your MCP client.

§3 — Verify the install

Your agent now has access to two tools and two prompts. Try this query:

"Find Azure RAG accelerators with eval coverage"

The agent should call `orchard.search` with parameters like `query="rag"` and `where="variety:azure AND trust_badge:eval_proven"` and render a Markdown table inline.

§4 — Next steps

§5 — Get help

The OrchardBrowse catalogCLI quickstartEdit on GitHub