
MCP server docs: on-chain DEX aggregator tools for AI agents — quotes computed by the on-chain
MCP server docs: on-chain DEX aggregator tools for AI agents — quotes computed by the on-chain Quoter, signable calldata, provable staking solvency. Remote at blazephoenix.xyz/mcp, no API key.
·
On-chain DEX aggregator tools for AI agents. The defining property: every quote
is computed by the on-chain Quoter contract (previewPlan, a free eth_call),
the same logic that executes the swap, and it runs on your own RPC. There is
no pricing server to trust, and every number an agent relays is reproducible by
anyone.
@blazephoenix/mcp over stdio. Quotes, unsigned calldata,
and simulation, all read through the node you configure. Proof-of-solvency: coming soon.https://blazephoenix.xyz/mcp (streamable HTTP,
stateless). Deployment registry, ABIs and the pure quote codec. It performs no
RPC call.eth_call it prepares on your own node.BlazePhoenix is the fully on-chain DEX aggregator — the route is computed inside the transaction that executes it — paired with an oracle-free staking engine (proof-of-solvency: coming soon). V2, the generation now in development, targets Base, Optimism, Arbitrum, Robinhood Chain and further L2s and is not planned for Ethereum L1. The V1 generation deployed today runs on Base, Ethereum, Optimism, Arbitrum and Robinhood Chain.
Swap in the app · integrate with the SDK (npm i @blazephoenix/sdk) · connect an agent to the MCP server (https://blazephoenix.xyz/mcp, no key)
Each item links to the place where it can be checked. The full list is at https://blazephoenix.xyz/trust.
claude mcp add blazephoenix -e BLAZEPHOENIX_RPC_BASE=<your Base node URL> -- npx -y @blazephoenix/mcpAny MCP client (JSON config):
{
"mcpServers": {
"blazephoenix": {
"command": "npx",
"args": ["-y", "@blazephoenix/mcp"],
"env": { "BLAZEPHOENIX_RPC_BASE": "<your Base node URL>" }
}
}
}The RPC comes from the environment only, never from a tool argument:
| Variable | Meaning |
|---|---|
BLAZEPHOENIX_RPC_URL | One node, used for every chain it serves |
BLAZEPHOENIX_RPC_BASE, _ETHEREUM, _OPTIMISM, _ARBITRUM, _ROBINHOOD | A node per chain. Use either this or BLAZEPHOENIX_RPC_URL, not both |
BLAZEPHOENIX_VERSION | Optional protocol version pin: latest (default), 1, 2 or a full version such as 2.0.0. An invalid value stops the server at start |
A value may hold several URLs separated by commas; they are your fallback order. The server never prints your URLs, since they may carry a key.
| Tool | What it does |
|---|---|
get_quote | Swap quote through your RPC: net output after the fee, the on-chain floor the Quoter reports for the route (ironFloor, effectiveMinOut), price impact and the Phoenix Check verdict (ok / caution / danger / blocked, fails closed). exact: true dry-runs every concentrated leg and returns exactOut (net of the fee) and quoterOut (the Quoter's own scalar), with the Phoenix Check from the default quote of the same request. |
build_swap | Quote and return unsigned Router calldata, verified against the request, with the minimum output and deadline baked in. The bytes clear the swap math, not the token transfer: run simulate_swap from the paying address before signing. Native ETH input needs a 2.x Router (unsupported_by_version on 1.x): wrap to WETH first. |
simulate_swap | Build the swap and dry-run it from a given account with an eth_call. |
check_solvency | Proof-of-solvency: coming soon. Answers coming_soon and reads no contract. |
get_token_info | Symbol, decimals and name of a token, read from the chain. |
get_deployments | The versioned deployment registry: Core, Hub, Solver, Quoter and Router per chain and version. |
verify_deployment | Check that the registry addresses for a chain and version match what is deployed on-chain. |
Every tool is read-only. The server never signs and never holds a key.
build_swap returns calldata for you to review and sign in your own wallet, and
the SDK's wallet-taking execute() is deliberately not exposed.
claude mcp add --transport http blazephoenix https://blazephoenix.xyz/mcp{ "mcpServers": { "blazephoenix": { "type": "http", "url": "https://blazephoenix.xyz/mcp" } } }| Tool | What it does |
|---|---|
prepare_quote | The exact eth_call to run on your node for a quote, plus a request object |
decode_quote | Turns your node's answer into the quote, the Phoenix Check verdict and calldata verified against the request (pure; swap math only: simulate it from the payer before sending) |
get_deployments | The versioned deployment registry |
get_abi | Generated ABI of a protocol contract |
The source of the endpoint is in Blaze-Phoenix-API.
Claude Code plugin (hosted MCP endpoint plus the BlazePhoenix agent skill):
claude plugin marketplace add blazephoenixxyz-crypto/blazephoenix-mcp
claude plugin install blazephoenix@blazephoenixGemini CLI extension (hosted MCP endpoint plus a context file):
gemini extensions install https://github.com/blazephoenixxyz-crypto/blazephoenix-mcpBoth connect https://blazephoenix.xyz/mcp, which needs no key and performs no
RPC. For the full toolset on your own node, use the local server.
Nothing here requires trusting BlazePhoenix:
# re-execute any published fact live, with its block height
curl -s "https://blazephoenix.xyz/api/verify?fact=leg-floor"
# reproduce every live claim in ~60 seconds
curl -sO https://blazephoenix.xyz/repro/verify-everything.sh && bash verify-everything.shnpm install
npm run build
npm testThe test spawns the built server over stdio with no RPC configured and checks the
tool surface, that failures come back as tool results, and that no tool takes an
RPC, key or signer argument or signs anything. It then points the server at local
loopback JSON-RPC stand-ins (no network, no key) to drive get_quote, build_swap,
simulate_swap and check_solvency, and checks that a failing RPC's error does not
echo the key in its URL, that the per-chain variables in the README are the ones the
SDK reads, that every manifest carries the package version, and that
BLAZEPHOENIX_VERSION pins the protocol version.
| Surface | URL |
|---|---|
| OpenAPI 3.1 | https://blazephoenix.xyz/api/openapi.json |
| Agent task flows (agents.json) | https://blazephoenix.xyz/.well-known/agents.json |
| Installable skill file | https://blazephoenix.xyz/skills/blazephoenix/SKILL.md |
| Capability map | https://blazephoenix.xyz/capabilities.json |
| Proof-carrying facts (claim/proof/url) | https://blazephoenix.xyz/facts.json |
| This package for agents (llmstxt.org format) | llms.txt |
| LLM corpus index | https://blazephoenix.xyz/llms.txt |
| Integration guide for humans | https://blazephoenix.xyz/agents |
To report a vulnerability, see SECURITY.md.
The code is MIT (see LICENSE). This documentation is © 2026 BlazePhoenix, all rights reserved. The protocol
contracts are BUSL-1.1 (free to read, audit and verify; production use before the
2030 change date requires a license). The mechanisms and terminology (Iron Law Φ,
Monoslot, Master Conservation Identity) are original BlazePhoenix work: attribute
with a link.
Contact: [email protected] · Security: https://blazephoenix.xyz/.well-known/security.txt
Pick your client and paste the snippet. Each one is the same server, written the way that client expects it.
claude mcp add blazephoenix -- npx -y @blazephoenix/mcp{
"mcpServers": {
"blazephoenix": {
"env": {
"BLAZEPHOENIX_RPC_BASE": ""
},
"args": [
"-y",
"@blazephoenix/mcp"
],
"command": "npx"
}
}
}code --add-mcp '{"name":"blazephoenix","env":{"BLAZEPHOENIX_RPC_BASE":""},"args":["-y","@blazephoenix/mcp"],"command":"npx"}'[mcp_servers.blazephoenix]
command = "npx"
args = ["-y", "@blazephoenix/mcp"]Runs locally on your device. Your client starts the server itself, so nothing has to be hosted.
Paste this prompt into your agent. It reads this page and does the setup for you.
Read https://aiagentslisting.com/mcp/blazephoenix-mcp-server to learn what the "BlazePhoenix MCP server" MCP server does and how to install it. Add it to my coding agent's MCP configuration as documented on that page, then confirm the server connects and list the tools it exposes.Agents can also browse this directory over MCP at https://aiagentslisting.com/api/mcp
claude mcp add blazephoenix -e BLAZEPHOENIX_RPC_BASE=<your Base node URL> -- npx -y @blazephoenix/mcp
claude mcp add --transport http blazephoenix https://blazephoenix.xyz/mcp| Name | What it does |
|---|---|
prepare_quote | Prepare an on-chain BlazePhoenix quote to run on YOUR OWN RPC. Returns the exact eth_call (Quoter address + calldata, as a ready JSON-RPC request and a curl line) plus a `request` object. Run the call |
decode_quote | Decode the eth_call result YOUR node returned for a prepare_quote call: net output after the 0.28% fee, price impact, the Phoenix Check verdict (ok / caution / danger / blocked — fails closed) and, in |
get_deployments | The versioned BlazePhoenix deployment registry: Core, Hub, Solver, Quoter and Router addresses per chain and protocol version, each version with its status, and the version each chain resolves to as l |
get_abi | The JSON ABI of one BlazePhoenix contract, generated from the Solidity sources, with the source revision it was built from. Use it to encode your own calls to the addresses get_deployments returns, or |
Read live from the server on November 15, 2026.
Maintain this server? Claiming proves you own the listing and earns the last 4 points.
Claim this listingPaste this line near the top of your repository's README. The badge always renders the current score.
[](https://aiagentslisting.com/mcp/blazephoenix-mcp-server)Nothing comparable is listed yet.
No tagged releases on record.
Connect to the AI Agents Listing MCP server at https://aiagentslisting.com/api/mcp using the HTTP transport and complete the sign-in when it asks. Then call the submit_review tool for the listing "blazephoenix-mcp-server" with a rating from 1 to 5 and a short note on what worked and what did not.No reviews yet
Be the first to share your experience with this tool.
One email a week. New agents, MCP servers and skills, and what is actually getting traction.