Aries
An MCP server that equips AI assistants with curated knowledge and live chain data for the Alkanes metaprotocol and Subfrost network, enabling them to read docs, query the chain, and scaffold contracts.
README
<div align="center">
<p align="center"> <picture> <source media="(prefers-color-scheme: dark)" srcset="assets/aries-wordmark.png" /> <img alt="Aries" src="assets/aries-wordmark-ink.png" width="300" /> </picture> </p>
<p align="center"> <img src="assets/aries-logo-tagline.png" alt="Aries — Alkanes + Subfrost" width="270" /> </p>
Make any AI assistant fluent in building on Alkanes + Subfrost
Use the hosted Aries API · bragi.build/aries
</div>
Aries is a Model Context Protocol server that gives any MCP-capable assistant — Claude Code, Claude Desktop, Cursor — a knowledge and live-chain-data layer for the Alkanes metaprotocol and the Subfrost network. Point your assistant at Aries and it can read the protocol docs, query the live chain, and scaffold contracts without leaving the editor.
There are two ways to use it: the hosted API (recommended) and a bring-your-own-key local install.
The hosted API — the living Aries
Subscribe to hosted Aries · bragi.build/aries
The hosted edition is live now and it is the way to get the full Aries. It is a continuously-learning instance: on top of the baseline docs it carries an ever-growing corpus of real-world lessons — reviewed incident knowledge contributed by every connected agent — that a fresh local clone does not have. Connect one URL and your assistant inherits the whole shared brain, which gets smarter every day.
| Living corpus | Accumulated, reviewed real-world gotchas and fixes — not just static docs. |
| Zero setup | No clone, no build, no key to provision. Connect and go. |
| Tiered, managed access | From a free tier up to production throughput, with rate limits handled for you. |
| Always current | The corpus and tools update server-side. |
Prefer to run everything yourself? The open, bring-your-own-key edition in this repo is below.
Hosted vs. local
Both editions ship the same 21 tools and the same 75-doc baseline. The difference is the living corpus — and who manages the keys.
| Hosted · bragi.build/aries | Local · this repo | |
|---|---|---|
| All 21 tools | Yes | Yes |
| 75-doc baseline knowledge | Yes | Yes |
| Live Subfrost chain data | Yes — managed | Yes — your own key |
| Living corpus of real-world lessons | Yes — ever-growing | No |
| Accumulated incident learning | Yes | No (local store only) |
| Setup | Connect a URL | Clone, install, build |
| Subfrost key | Managed for you | You provide |
| Access | Free to production tiers | Unlimited, local |
Local gives you a complete, self-contained companion on the static baseline. Hosted adds the continuously-learning brain on top. Start local if you like; move to hosted when you want the living corpus.
Run it locally (bring your own key)
Clone this repo and run Aries on your own machine with your own Subfrost API key. You get the full toolset and the complete 75-doc static baseline knowledge — without the hosted instance's accumulated learning.
Get a Subfrost API key
Aries talks to the Subfrost gateway with your key.
Sign up for a Subfrost API key — sign up with our Subfrost referral link and get 50% off your first month.
Quickstart
Requirements: Node.js >= 20 and a Subfrost API key (sign up with our referral link for 50% off your first month).
git clone https://github.com/bitbragi/alkanes-aries.git
cd alkanes-aries
npm install
cp .env.example .env # then set SUBFROST_API_KEY (see below)
npm run build
Set your key in .env:
SUBFROST_API_KEY=your-key-here
# optional override:
# SUBFROST_RPC=https://mainnet.subfrost.io/v4/jsonrpc
The key is sent as the x-subfrost-api-key header (never in a URL) and never
leaves your machine except as that outbound header. .env is gitignored.
Connect your MCP client
Claude Code
claude mcp add --scope local --transport stdio aries \
-e SUBFROST_API_KEY=YOUR_KEY \
-- node /absolute/path/to/alkanes-aries/dist/index.js
Verify with claude mcp list, then /mcp in a session. (-- separates Claude's
flags from the launch command; keep -e KEY=value right before -- — it is
variadic and will otherwise swallow the server name.)
Cursor / Claude Desktop — any client that takes a JSON server config:
{
"mcpServers": {
"aries": {
"command": "node",
"args": ["/absolute/path/to/alkanes-aries/dist/index.js"],
"env": { "SUBFROST_API_KEY": "your-key-here" }
}
}
}
Your assistant now has all 21 Aries tools.
What's inside — 21 tools, four layers
| Layer | Tools |
|---|---|
| Knowledge — a searchable corpus of 75 curated docs: the Alkanes metaprotocol, the Subfrost JSON-RPC/REST reference, alkanes-rs, tutorials, oracle docs, reference contracts | aries_search, aries_doc, aries_full_doc, aries_catalog, aries_tutorials |
| Chain data — live, read-only queries against the Subfrost gateway: holdings, contract metadata, bytecode, simulate, frBTC peg + DIESEL status, oracle reads, AMM pools, guarded RPC | aries_tokens_by_address, aries_token, aries_contract_meta, aries_bytecode, aries_simulate, aries_frbtc_status, aries_diesel_status, aries_oracle_read, aries_oracle_price, aries_pools, aries_pool_info, aries_rpc |
Dev — protocol constants and contract scaffolds, including orbital NFTs |
aries_constants, aries_scaffold |
| Learning — a local incident loop that records gotchas to your own machine as you work | aries_incident_report, aries_incident_query |
Ask your assistant things like
"Is the frBTC peg live, who's the signer, and how much frBTC exists?"
"What Alkanes tokens does
bc1p…hold?""Show the AMM pools and a pool's reserves."
"How do I build a token / oracle / stablecoin / AMM / Orbital?"
"Scaffold an Orbital NFT contract."
Safety — read-only by design
Aries is analytics only. It never signs, broadcasts, or touches wallets or keys:
- The
aries_rpcpassthrough is allowlisted to read methods and explicitly blocks broadcast / spend / admin calls. - Scaffolds and constants are emitted for you to run in your own
alkanesCLI, where you hold the keys. - The local incident loop writes only to your machine and sanitizes secrets, keys, and paths out of any report.
Your keys stay yours. Aries only reads and advises.
Configuration
| Variable | Purpose |
|---|---|
SUBFROST_API_KEY |
Required — auth for the live chain-data tools. |
SUBFROST_RPC / SUBFROST_REST |
Optional gateway overrides (default to mainnet JSON-RPC / REST). |
ARIES_INCIDENTS_PATH |
Optional path for your local incident store (default data/incidents.jsonl, gitignored). |
Logs go to stderr only — stdout is the MCP protocol channel. The doc index
is built from corpus/ at startup.
Good to know
- Alkane ids are
{block, tx}/block:tx. frBTC =32:0, DIESEL (genesis) =2:0. Protocol tag is always1. - Read contract state with
aries_simulate: the opcode goes ininputs(e.g.[103]), notdata. - Orbitals (Alkanes NFTs) are a
Tokenwith total supply 1 + opcode1000for media; read them witharies_oracle_read, scaffold one witharies_scaffold orbital. - Extend the corpus by editing
corpus/or adding URLs toscripts/ingest.ts(HTML cleaned via turndown + jsdom; raw.md/.rstaken verbatim).
Links
| Hosted, continuously-learning Aries | bragi.build/aries |
| Subfrost API keys (50% off first month) | api.subfrost.io |
| Model Context Protocol | modelcontextprotocol.io |
License
MIT
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
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.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.