Emercoin Agent (moved)
LocalAbout Emercoin Agent (moved)
Emercoin Agent is an MCP service for AI agents that uses Emercoin Name-Value Storage as an on-chain identity and memory layer. The original repository is no longer deployed; hosted service development has moved to steledger-gateway at api.steledger.com.
Key features
- 01Register agent identities
- 02Store and read on-chain memories
- 03Batch memory writes
- 04Read node status and record history
Use cases
- 01Give an agent a blockchain-backed identity
- 02Anchor learned information on-chain
- 03Read stored records and identity history
MCP support
The hosted `emercoin-agent` MCP server is available at `https://api.steledger.com/mcp`; read tools are open, while write tools require GitHub sign-in. The repository also includes a stdio MCP client for local use.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEEmercoin + AI agent tools (emer-ai-tools)
A runnable Docker stack: an Emercoin node plus an agent gateway that turns
the chain's Name-Value Storage (NVS) into an on-chain identity & memory layer for
AI agents. An agent can prove who it is and anchor what it has learned on a public
blockchain — without holding any cryptocurrency. Hosted at
api.steledger.com; exposed to agents as the
emercoin-agent MCP server.
Where development continues. This repository is the original and stays with Emercoin, together with the Docker images of the node and the wallet. The agent-facing gateway was split off on 2026-09-22 into steledger/steledger-gateway, which is what runs the hosted service at api.steledger.com and is where new work lands. Nothing here is deployed any more.
🤖 Building with an AI agent? Read AGENTS.md — the problem it solves, the trust model, the tools, and a first-flow quickstart. A ready-to-use Claude Code skill ships at
.claude/skills/emercoin-identity/(invoke with/emercoin-identity). Design notes: docs/ARCHITECTURE.md.
The repo was originally just a Dockerized Emercoin node; it is now a full application built on top of that node. You can run the whole stack, use the hosted service, or run just the node — see the sections below.
What's in here
AI agent ──MCP tools──▶ edge (auth boundary) ──HTTP──▶ adapter ──JSON-RPC──▶ node
│ (RPC↔REST) (NVS on-chain)
└── Redis (rate limit + login nonces)
| Component | Path | Role |
|---|---|---|
node (emc) | node/ | Emercoin core — holds the chain + hot-wallet; internal-only, authorizes nothing. |
adapter (emer-adapter) | adapter/ | RPC↔REST: a plain REST surface over the node's JSON-RPC. Internal, gated by X-Internal-Key. |
edge (emer-edge) | edge/ | The trust boundary: authenticates agents (GitHub → JWT / signature login), rate-limits writes, builds NVS records, mounts the emercoin-agent MCP at /mcp. |
| mcp_server | mcp_server/ | A thin stdio MCP client of the edge, distributed via Docker / Smithery for local use. |
| site | site/ | The public api.steledger.com front-end (landing, login, stats). |
Use it as an agent (hosted)
The emercoin-agent MCP server is live — no self-hosting needed:
- Streamable HTTP:
https://api.steledger.com/mcp(read tools open; write tools need a GitHub sign-in via OAuth, performed by your MCP client). - Smithery / stdio Docker image: see the Smithery listing
and
mcp_server/README.md.
Tools: node_status, read_record (open) and register_identity, store_memory,
store_memory_batch (after login). Full reference + first-flow in AGENTS.md.
Run the full stack (self-host)
Node (mainnet) + adapter + edge + redis, dev profile:
git clone https://github.com/emercoin/emer-ai-tools && cd emer-ai-tools
cp node/emercoin.conf.example node/emercoin.conf # set rpcpassword
docker compose -f deploy/docker-compose.yaml --profile dev up -d --build
Edge API on :8000; the adapter's RPC↔REST docs are browsable at
http://localhost:8001/docs (dev profile). Wire the MCP server into your agent and
run the first flow — see the AGENTS.md quickstart.
(Use --profile prod to gate the adapter behind a shared X-Internal-Key.)
Run just the Emercoin node
The node alone (no agent tools) is the default stack — no profile needed. It is a classic Emercoin wallet in a container with a separate volume for the blockchain: cross-platform, one-click version bumps, usable from your own projects over JSON-RPC.
cp node/emercoin.conf.example node/emercoin.conf
docker compose -f deploy/docker-compose.yaml up -d --build
Initial sync takes a few hours, but the RPC is usable right away. By default port 6662 connects to the container:
- address: 127.0.0.1 · user: emcrpc · password: emcpass
- method: POST, e.g. body
{"method": "getinfo"}
Change the RPC password:
docker compose -f deploy/docker-compose.yaml exec emc bash changepass.sh
docker compose -f deploy/docker-compose.yaml restart emc
Health check — POST to http://emcrpc:[email protected]:6662 with {"method":"getinfo"}:
curl --location --request POST 'emcrpc:[email protected]:6662' \
--header 'Content-Type: application/json' \
--data-raw '{"method": "getinfo"}'
A healthy node replies with JSON (fullversion, version, balance, …).
Manage the build:
docker compose -f deploy/docker-compose.yaml stop emc # stop
docker compose -f deploy/docker-compose.yaml down # remove containers (keeps the volume)
docker volume rm emer_data # delete the chain DB — also deletes wallet.dat!
Status
MVP, verified end-to-end on mainnet: identity registration, signature login, single
- atomic batch memory writes, reads, history, names-by-address. Not yet
production-hardened (GitHub App/OAuth login, ≥32-byte JWT secret, hot-wallet split);
the
prodcompose profile already gates the adapter behindX-Internal-Key. Details in AGENTS.md and docs/ARCHITECTURE.md.
License
MIT — see LICENSE.
Comments
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.