# Sally Skills

> Sally Skills is a metered MCP server + REST API that exposes the clinical-grade metabolic-health intelligence behind Sally AI (the AI inside the A1C Insights iOS app) to AI agents. One bearer key (`sk-sally-…`), six skills, per-call pricing in USD, no subscription.

Canonical names:
- **Sally AI** — the underlying metabolic-health intelligence (the brand).
- **Sally Skills** — the MCP + REST API gateway (this product, at `sally.a1c.io`).
- **A1C Console** — developer dashboard (this site, at `console.a1c.io`).
- **A1C Insights** — consumer iOS app, the identity source for every Sally Skills API key.

## What an agent can do with Sally Skills

- `health_sync` (FREE) — pull 64+ daily biomarkers from the user's wearables, CGM, sleep, vitals, activity, and environment into the agent's context. Requires the user has the A1C Insights app installed.
- `chat_with_sally` ($0.003 / call) — conversational access to Sally's preventive-health and Traditional Chinese Medicine knowledge with source citations. Evidence-graded, never invented.
- `analyze_lab_result` ($0.008 / call) — parse and interpret lipid, HbA1c, CBC, thyroid, hormone, micronutrient panels via VLM OCR. Returns risk flags, reference ranges, and ADA / ACC / ESC guideline context.
- `food_journal` ($0.004 / call) — grade a meal photo. Returns macros, glucose-spike prediction, and Smart-vs-Trap categorisation.
- `health_insights` ($0.003 / call) — morning / afternoon / evening readouts synthesised from sleep, vitals, and activity. Built for daily-briefing agents.
- `metabolic_overview` ($0.005 / call) — full CGM snapshot for a date: time-in-range, glycemic variability, dawn phenomenon, postprandial response curves. Requires the user has an active CGM in A1C Insights.

## Docs

- [Landing](https://console.a1c.io/): product overview, hero, pricing.
- [Docs](https://console.a1c.io/docs.html): introduction, quickstart, authentication, MCP integration (Claude Code, Claude Desktop, Manus, Perplexity, Codex, Windsurf, Google Antigravity), REST protocol, wallet, error codes.
- [MCP integration section](https://console.a1c.io/docs.html#mcp): copy-paste config for every supported agent.
- [Skill catalog](https://github.com/Sally-A1C/ai-sally-skills/blob/main/SKILLS.md): full input + output schemas with error codes.
- [System design](https://github.com/Sally-A1C/ai-sally-skills/blob/main/SYSTEM_DESIGN.md): architecture, sequence diagrams, billing model.
- [Per-skill mechanics](https://github.com/Sally-A1C/ai-sally-skills/tree/main/packages): one README per skill.

## How to call Sally Skills

Server endpoint: `https://sally.a1c.io`
Auth: `Authorization: Bearer sk-sally-...` (mint at https://console.a1c.io/keys)
Transports: MCP (Streamable HTTP, JSON-RPC 2.0) at `/mcp`, REST at `/v1/call` and `/v1/skills/:name`.

Universal MCP config (works in Claude Code, Claude Desktop, Cursor, Manus, Perplexity, Codex, Windsurf, Google Antigravity):

```json
{
  "mcpServers": {
    "sally": {
      "url": "https://sally.a1c.io/mcp",
      "headers": { "Authorization": "Bearer sk-sally-..." }
    }
  }
}
```

REST smoke test:

```bash
curl -sS https://sally.a1c.io/v1/call \
  -H "Authorization: Bearer sk-sally-..." \
  -H "Content-Type: application/json" \
  -d '{"skill":"health_sync","input":{}}' | jq
```

## Optional

- [GitHub: gateway source](https://github.com/Sally-A1C/ai-sally-skills): metered Hono gateway in TypeScript/Bun.
- [GitHub: public docs + agent guides](https://github.com/a1c-ai-agent/sally-skills): Markdown docs site + `SKILL.md` auto-routing layer for agents.
- [A1C Insights on the App Store](https://apps.apple.com/id/app/a1c-insights/id6748399956): the iOS app users install to mint an account.
- Contact: ai@sallya1c.com