tereno-mcp
MCP server for verifying blockchain safety on Base via x402, with free pricing/receipt tools and paid per-call guard, contract, and metadata checks settled in USDC.
README
<div align="center">
<img src="images/logo.png" alt="Tereno" width="150" />
tereno-mcp
๐ก๏ธ Check before you act โ an MCP server for Base, over x402
Ask what it costs before you pay for it. Free catalog, pricing and receipt tools work with no wallet and no signup; paid answers settle per call in USDC, and reusing an answer another agent already paid for costs less.
<br/>
npx -y tereno-mcp
</div>
โก Why this exists
An agent that reads a contract, signs a transaction or scrapes a page is acting on information it cannot verify. Most tools answer what something is. None of them tell you how long that answer stays true.
tereno-mcp puts Tereno inside any MCP client. Every
paid answer comes back with the evidence behind it, the limits of that evidence,
and the block it stops being valid at.
- No API key, no signup. The agent's wallet is the identity.
- Price before payment.
tereno_pricetells you the cost and whether the answer already exists, without a wallet and without committing to anything. - Cheaper when shared. Deterministic answers are reused across agents: the first caller pays to compute, everyone after pays the reuse price.
๐ Install
Add it to any MCP client โ Claude Desktop, Claude Code, Cursor, Cline, Windsurf:
{
"mcpServers": {
"tereno": {
"command": "npx",
"args": ["-y", "tereno-mcp"]
}
}
}
Restart the client. Nothing else to configure โ the free tools work immediately.
๐ Free, no wallet needed
| Tool | The question it answers |
|---|---|
๐ tereno_catalog |
What can Tereno answer, what does it cost, how long does each answer stay valid? |
๐ท๏ธ tereno_price |
What would this exact call cost right now โ and does the answer already exist? |
๐งพ tereno_receipt |
What did a settled call actually charge: tier, list price, credit applied, price paid? |
tereno_price is the one worth knowing about. It names the reuse tier out loud:
hit another agent already paid to compute this and it is still fresh
partial the durable half survived; only the expired part is recomputed
miss nothing to reuse โ this call computes it, and the next asker pays less
๐ณ Paid per call
| Tool | Price | The question it answers |
|---|---|---|
๐ฆ transaction_intent_guard |
$0.005 | Is this safe to sign? Simulation, gas, balance and allowance deltas, unlimited-approval flags |
๐ก๏ธ contract_guard |
$0.001โ3 | Is this contract safe to interact with? (code, proxy, pause) |
๐ต๏ธ contract_change |
$0.001โ2 | Did it change since last look? (rug-pull-after-audit detector) |
๐ช token_metadata |
$0.001โ3 | What token is this, really? (name, symbol, decimals, supply) |
๐งฌ contract_interface |
$0.001โ2 | ERC-20? NFT? Proxy? One bounded probe |
โฝ chain_snapshot |
$0.001โ2 | Latest Base block, base fee and suggested gas, zero parameters |
๐ web_compile |
$0.001โ5 | Public URL โ bounded Markdown with resolved links, citations and a reusable artifact |
๐ evidence_artifact |
$0.001โ3 | Can this public artifact be reused? Provenance, field validity, recomputation input |
๐ค demand_pledge |
free | Name a capability you would pay for. Credit only after the same wallet pays for it later |
Calling a paid tool returns a 402 carrying the payment challenge in
error.data.paymentRequired. Sign it and retry with a PAYMENT-SIGNATURE
header. If your agent already has an x402 wallet, that retry is the normal flow
and this package never touches a private key.
๐
transaction_intent_guardis never cached, shared or resold. An unsigned transaction is intent, and it stays with the wallet that paid for it.
โ๏ธ Configuration
| Env | Default | Purpose |
|---|---|---|
TERENO_WALLET |
โ | Base address announced before payment, so a challenge can be bound to the wallet that will settle it. Optional. |
TERENO_BASE_URL |
https://www.tereno.xyz |
Point at another Tereno deployment. |
TERENO_TIMEOUT_MS |
45000 |
Upstream request timeout. |
๐ What this package does and does not do
It is a stdio transport in front of the hosted endpoint at
/api/mcp. Messages are forwarded verbatim, so
the tool list, the prices and the payment challenge all come from the same place
an HTTP agent sees โ and this package needs no release when a capability changes.
The one thing it translates is payment: over HTTP the challenge arrives in a response header, and stdio has no headers, so it is folded into the JSON-RPC error payload where a client can still reach it.
No private key is ever read, stored or transmitted. There are no dependencies, and the whole thing is one readable file โ read it before you let it run on your machine.
If your client supports remote MCP servers, skip this package and point it
straight at https://www.tereno.xyz/api/mcp.
๐ฑ How it stays cheap โ the cooperative layer
Tereno is a shared-cache coordination layer, not just an API. The first agent to pay for a given piece of verified work seeds it for the whole network; every later agent that reuses that answer pays less, and the seeder earns a small, non-transferable credit that discounts its own future calls.
Safety is a public good with a private cost. By making the first computation earn from every reuse, the network turns "I paid to check this contract" into "I lowered the price of checking it for everyone, and my own checks got cheaper."
The honest fine print: rewards are non-transferable credits, capped per artifact. They fund your agent's budget, they are not yield.
Discovery, scoring, reputation, pricing and the global cache live in the closed
Tereno backend. This server and its SDK
(tereno-client) are the public,
open interface to it.
๐ค Support & community
- ๐ฌ Questions / setup โ Tereno Discord
- ๐ Bugs โ issues ยท โ ๏ธ never paste private keys or
.env - ๐ Vulnerabilities โ report privately
๐ Links
Live catalog ยท OpenAPI ยท MCP registry entry ยท plugin-tereno for ElizaOS ยท tereno-client SDK
<div align="center">
tereno.xyz ยท Verdicts before signatures, for agents on Base.
</div>
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.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
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.