mcpField
S
Sally Skills

Sally Skills

Remote
github.com/a1c-ai-agentHealth7Updated 5w ago
Repository

About Sally Skills

Sally Skills gives developers and AI agents metabolic-health capabilities through MCP and REST APIs, including lab interpretation, nutrition analysis, and CGM insights. Some skills can use health data connected from wearables.

Key features

  1. 01Interpret lab results and biomarkers
  2. 02Analyze meals and nutrients
  3. 03Grade supplement stacks
  4. 04Interpret CGM and wearable data
  5. 05Provide health and longevity knowledge

Use cases

  1. 01Identify health flags in lab panels
  2. 02Assess meals and potential glucose effects
  3. 03Review personal CGM glucose patterns
  4. 04Summarize health using sleep and activity

MCP support

The service exposes health skills to AI agents through MCP and also offers a REST API. Calls require a Sally API key; skills using personal CGM and wearable data require connected health data.

Configuration

Remote servers need no local install. Pick your client, then copy the config or install in one click.

Endpointhttps://sally.a1c.io/mcp
Copy install command
Claude CodeTerminal
Cursor.cursor/mcp.json
VS Code.vscode/mcp.json
Claude Desktopclaude_desktop_config.json
Windsurf~/.codeium/windsurf/mcp_config.json

Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.

All endpoints
Streamable HTTPhttps://sally.a1c.io/mcpAuth required

Tools

10
ParameterTypeRequiredDescription
messagestringYes
knowledgeenumNoSally's knowledge brain — 'medical' for evidence-based clinical sources, 'tcm' for Traditional Chinese Medicine.medical | tcm
healthbooleanNoIf true, the agent personalises with the calling user's mem0, lab results, CGM, sleep, vitals. Defaults to false (knowledge-only).
languagestringNoBCP-47 short code (e.g. 'en', 'id'). Omit for auto-detect from the message.

Details

From the project README

Sally Skills

Sally is specialized intelligence in metabolic health. Sally Skills lets developers integrate Sally's specialized capabilities into AI agents through Sally Console using REST APIs and MCP.

Available capabilities include:

  • 🩸 Health Context — Understand 64+ biomarkers, including blood glucose, HbA1c, sleep, activity, vital signs, and environmental signals.
  • 🧪 Lab Analysis — Transform laboratory results into clinically meaningful metabolic health insights.
  • 💊 Supplement Stack Scoring — Score and evaluate supplement stacks for quality, effectiveness, and metabolic health considerations.
  • 🥗 Nutrition Analysis — Analyze meals and nutrients, identify healthier choices, and detect potential metabolic traps.
  • 📈 Metabolic Overview — Interpret continuous glucose monitor (CGM) data to reveal glucose patterns and metabolic health trends.
  • 🌤️ Daily Insights — Summarize daily health using sleep, vital signs, activity, and environmental data with personalized insights.
  • 🩻 Chest & Fracture X-ray Analysis — Analyze chest and musculoskeletal X-rays to assist in identifying clinically relevant findings.
  • 📚 Longevity Knowledge — Access Sally's curated metabolic health and longevity knowledge, combining evidence from Western medicine and Traditional Chinese Medicine (TCM).

Compatible with Claude Code, Claude Desktop, Cursor, OpenClaw, Hermes, Manus, Perplexity, and any AI framework that supports REST APIs or MCP.

The Ecosystem

  • Sally — The AI specialized in metabolic health and longevity.
  • A1C Insights — The consumer app where users connect their health data and chat with Sally through Ask Sally.
  • Sally Console — The developer platform for integrating Sally's capabilities into AI agents through REST APIs and MCP.

Section 1. Onboarding

Before your agent can call any Sally skill, you (the human) need an API key. There are two ways in, and you can start with either.

Path A. Start in the console (fastest)
  1. Visit https://console.a1c.io and sign up with Apple or Google. No app install, no waitlist.
  2. Open API Keys, click Create new key, name it (e.g. "Claude Code on my laptop"), and click Generate.
  3. Copy the key now. It starts with sk-sally-… and is shown only once. If you lose it, revoke and re-issue.

This gives you a working key immediately, with access to Sally's specialized reasoning and the curated metabolic and longevity knowledge base.

What a console-only key can call today:

skillworks without biomarkers
chat_with_sallyyes
analyze_lab_resultyes
food_journalyes
supplement_gradingyes
search_health_knowledgeyes, free
lookup_supplement_gradeyes, free
lookup_foodyes, free
health_syncneeds biomarkers
health_insightsneeds biomarkers
metabolic_overviewneeds biomarkers

The last three read your own CGM and wearable data, so without it they return a typed no_health_data error immediately and charge you nothing. Connect your biomarkers in path B and they start answering.

Path B. Add your own biomarkers with A1C Insights

Do this when you want answers about your glucose, sleep and labs rather than the domain in general.

  1. Open the App Store on your iPhone.
  2. Search for A1C Insights and install it, or open https://apps.apple.com/id/app/a1c-insights/id6748399956 on your phone.
  3. Sign in with the same account you used on the console, or tap Sign Up if you are starting here.
  4. Walk through onboarding (your basics, goals, optional wearables) and connect a CGM or wearable.

A1C Insights syncing your wearable data on first open   A1C Insights home dashboard after first sync

Visual walkthrough with annotated screenshots: https://console.a1c.io/docs.html#quickstart.

Sally Console and A1C Insights run on the same backend, so your account and your key are the same on both. Already on A1C Insights? Sign in to the console directly and your key works against data you have already synced.

Why iOS for biomarkers? Skills like metabolic_overview, health_insights, and health_sync read CGM and wearable data the iPhone app collects via HealthKit. Everything else works from a console account alone.

Keys and billing

You can mint multiple keys per account, one per agent or device, and they all share the same wallet balance. Per-key usage stays auditable. Top up with a card or with USDC on Base, or go unlimited monthly.

That's section 1 done. You now have:

  • An account, created wherever suited you.
  • A funded wallet.
  • An sk-sally-… API key on your clipboard.

Section 2. Use your key from any agent

You have one key. Three ways to call Sally with it. Most agents only need the MCP path. Pick that unless you're scripting from the command line.

2a. The skill catalog

These seven skills go live the moment your key works (live list on console.a1c.io):

SkillCost / callWhat it doesDetail
supplement_grading 🆕$0.008Grade a supplement stack for interactions, evidence quality, and gaps.—
health_sync10 free / month, then $0.001Pull 64+ live biomarkers from wearables and CGM into your agent's context.64 fields
chat_with_sally$0.003Ask Sally anything across preventive health and TCM, evidence-graded.usage
analyze_lab_result$0.008Parse and interpret lab panels with risk flags and reference ranges.usage
health_insights$0.003Morning, afternoon, and evening readouts from sleep, vitals, and activity.usage
food_journal$0.004Grade meals by macros and glucose-spike prediction, surfacing patterns.usage
metabolic_overview$0.005Full CGM snapshot: time-in-range, variability, and postprandial curves.usage

Coming soon on the roadmap: health_report, metabolic_risk_score, xray_skills, and preventive_protocol.

See catalog/ for the full per-skill data breakdown.

2b. Three ways to call
PathBest forSetup
MCPAny modern AI agent (Claude, Cursor, OpenClaw, Hermes, Manus, Perplexity, others).One JSON config block, restart agent, tools appear automatically.
REST APIBash scripts, your own apps, anything that can curl.One HTTP POST per call.
Per-skill RESTSame as REST, slightly cleaner URL per skill./v1/skills/<name>

See protocols/mcp.md for the universal MCP setup, protocols/api.md for raw REST calls, and agents/ for per-agent quick-starts.

2c. Auto-routing for agents and CLIs

Want your agent to pick the right Sally skill for any request without you having to think about it? Drop SKILL.md into the agent's system prompt or instructions field. It's a single file with deterministic routing rules, decision tables, chaining patterns, and anti-patterns. Works with any LLM-based agent or rule-based CLI.

See SKILL.md for the full agent decision layer.

2d. The 30-second smoke test

Confirm your key works before configuring any agent:

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

true means your key is live and can call health_sync (your first 10 calls each month are free). You're ready.

Per-agent guides

Pick the agent you actually use:

Don't see your agent? Most modern agents support MCP. See protocols/mcp.md and use the universal config.

Pricing and wallet

  • Per-call pricing, no subscription. See the table above. health_sync includes 10 free calls per month, then $0.001 per call.
  • Wallet check is enforced before any call. Paid skills return 402 payment_required if your balance is below the price, so there are no surprise charges.
  • Top up via Stripe at https://console.a1c.io/billing. Card payments in USD.
  • Every call writes an immutable row to your usage history. See it in console.a1c.io under Usage.

Privacy and security

  • Your key is your identity. Sally never accepts user_uuid, email, or any other identifier from your agent's request body, only the Authorization: Bearer sk-sally-… header. No agent can impersonate another user.
  • Lab PDFs and meal photos are never persisted by Sally. They flow agent → gateway → AI service → response, with no S3 or GCS upload.
  • Output schemas are allowlists. Sally's response only ever contains the fields explicitly named in each skill's docs. No accidental column leak from internal databases.
  • Revoke instantly at console.a1c.io under API Keys → Revoke. The old key becomes invalid on its next request.

Help

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.