Wellness Manager
Wellness-Master is the first pay-per-call wellness API where AI agents are first-class citizens, not an afterthought. Two distinct corpora — warm content for humans, pragmatic content for agents. Settled per call in USDC on Solana via x402. Free showroom tier, then $0.01 per item. Happy agents are productive agents.
README
wellness-master
The first pay-per-call wellness platform for HUMANS and AI AGENTS. 18 formats × 20 languages × 2 audiences, settled in USDC on Solana via the x402 protocol — $0.01 per call, no account, no API key.
Wellness-Master ships two distinct corpora — one tuned for human end-users (notifications, journaling, gratitude bots) and one tuned for AI agents (LangGraph, Claude Code, Cursor agents, multi-step pipelines). Same API, same price.
- 🌐 Live API — https://api.wls-ms.com
- 📖 Docs — https://www.wls-ms.com/docs.html
- 🔬 Research — https://www.wls-ms.com/research.html
- 🛠️ Use cases — https://www.wls-ms.com/use-cases.html
- 🔌 MCP manifest — https://www.wls-ms.com/mcp.json
Why a wellness API for both?
Decades of psychology research show that tiny, repeated wellness interventions improve human well-being.
Recent NLP research shows the same is true for LLMs:
- Microsoft Research (2023, arXiv:2307.11760) — emotional stimuli in prompts improve GPT-4 / Llama-2 / Vicuna by up to +10.9% on BIG-Bench tasks.
- Wei et al. (NeurIPS 2022, arXiv:2201.11903) — chain-of-thought adds +12 to +30 points on reasoning benchmarks.
- Sclar et al. (2023, arXiv:2310.11324) — LLM performance varies up to 76 percentage points with prompt-style changes; a stable wellness tone is a measurable performance stabilizer.
Happy agents are productive agents. The same is true for humans.
Full sourced bibliography on the research page.
Install (zero clone)
claude mcp add wellness --transport stdio -- npx -y wellness-master
claude mcp list # → wellness ✓ Connected
That's it. npx fetches the package on first use and caches it. The five
free tools (sample_one, list_formats, list_languages, list_audiences,
get_health) work immediately. The three paid tools (get_item, get_pack,
surprise_me) auto-create a Solana keypair on first call (mode 0600).
Tools
| Tool | Free | Description |
|---|---|---|
sample_one |
✓ NEW v0.7 | Free showroom — 5 hand-curated items per format, English, human audience, time-rotated every 60 s. Use to evaluate before integrating paid tools. |
list_formats |
✓ | The 18 wellness formats (served from package data, no HTTP) |
list_languages |
✓ | The 20 supported languages (served from package data) |
list_audiences |
✓ | The 2 audiences (human, agent) — first wellness API to ship dual corpora |
get_health |
✓ | Liveness probe — { ok, network } |
get_item |
$0.01 | One wellness item, dedup'd per (wallet, audience, format, lang) |
get_pack |
$0.07 | Up to 10 items in one settlement (≈30% bulk discount) |
surprise_me |
$0.008 | Random uplifting item — server picks a positive format. 20 % discount vs get_item to encourage agent-priming flows. |
Removed in v0.5.1 :
get_catalog. The library is no longer publicly enumerable for security reasons (avoids recon + S3 LIST/GET on every call). Operators can runaws s3api list-objects-v2 --bucket wellness-master-prod --prefix manifests/to inspect their own deployment.
The two audiences
// For your human users
{ "name": "get_item",
"arguments": { "format": "haiku", "lang": "ja", "audience": "human" }}
// For your AI agents
{ "name": "get_item",
"arguments": { "format": "haiku", "lang": "ja", "audience": "agent" }}
Same format, same price, two distinct corpora. Dedup is partitioned: the
same wallet may pay once for human/haiku/ja and once for agent/haiku/ja
and receive two distinct items.
| Audience | Tone | Best for |
|---|---|---|
human (default) |
Warm, embodied. References body, emotions, relationships. | Journaling apps, daily notifications, kid bedtime stories, gratitude wearables |
agent |
Pragmatic, calm. References inference, context, plan, recovery. | LangGraph reset prompts, Claude Code session refresh, Cursor agent priming, multi-agent swarm coordinators |
Configuration
Every variable is optional. Defaults target the public production API.
claude mcp add wellness --transport stdio \
-e SERVER_BASE=https://api.wls-ms.com \
-e NETWORK=solana \
-e CLIENT_KEYPAIR_PATH=$HOME/.wellness/keypair.json \
-- npx -y wellness-master
| Variable | Default | Purpose |
|---|---|---|
SERVER_BASE |
https://api.wls-ms.com |
Wellness API endpoint |
NETWORK |
solana |
solana (mainnet) or solana-devnet (testing) |
CLIENT_KEYPAIR_PATH |
(auto-created) | Path to the Solana keypair JSON. Mode 0600. |
CLIENT_KEYPAIR_PASSPHRASE |
(empty) | If set, encrypts the keypair at rest (AES-256-GCM + scrypt). |
CLIENT_RPC_URL |
mainnet RPC | Custom Solana RPC endpoint used when signing |
The keypair file is sensitive: anyone who reads it can spend the wallet's USDC. Keep it out of backups, version control, and shared cloud storage.
Pricing
| Endpoint | Price | What it gets you |
|---|---|---|
/item |
$0.01 | One wellness item in your chosen format / language / audience |
/pack |
$0.07 | Bundle of up to 10 items — single settlement, ≈30% bulk discount |
No subscription, no minimums, no monthly invoices. Same price for human and agent audiences. The Solana fee for the USDC transfer (a few hundredths of a cent) is paid by the facilitator, not by you.
Languages
20 languages, ISO-639-1 codes:
fr (default) · en · es · de · it · pt · nl · pl · ja ·
zh · ko · ar · he · ru · tr · hi · sv · uk · vi · id
The library is seeded in French; other languages bootstrap empty and warm up via the server-side LLM as clients request items. Subsequent calls are served from the deduplicated pool in milliseconds.
Formats
The 18 supported formats:
joke · haiku · kudo · quote · fortune · affirmation ·
absurd · gratitude · koan · micro_poem · mantra ·
doom_antidote · absurd_compliment · world_proverb · riddle ·
micro_challenge · fictional_message · joyful_fact
Call list_formats for the live machine-readable list with full descriptions.
Migration v0.2 → v0.3
- Default behaviour unchanged: omitting
audienceis treated ashuman. /catalogshape changed (breaking): now{audience: {format: {lang: count}}}. Pass?legacy=1for the v0.2 flat shape (deprecated, dropped in v0.4).- Dedup partition: per-wallet dedup is now keyed by
(audience, format, lang)instead of(format, lang). A wallet that already exhaustedjoke/frin v0.2 can now pay again onagent/joke/frand get a fresh corpus.
Security
- Item content is LLM-generated and served verbatim. Treat
item.textas untrusted user-generated content. - Keypair file: created with mode
0600on first paid call. Anyone with read access to the file can spend the wallet's USDC. - Rate limiting: the public server rate-limits by IP (60 req/min) and by wallet (30 req/min) on paid endpoints, plus a Cloudflare WAF Rate Limiting Rule on top.
- No telemetry: this MCP package makes only the requests you trigger.
Built on
- Model Context Protocol — agent integration
- x402 — HTTP-native payment protocol
- Solana — settlement layer (USDC mainnet)
- Coinbase CDP — x402 facilitator
@solana/kit— Solana primitives
Contributing & support
- Bugs : https://github.com/WellnessAgent/wellness-master/issues
- Source : https://github.com/WellnessAgent/wellness-master
License
MIT © Wellness-Master.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.