Hedera's infrastructure is built. The last mile — getting live on-chain intelligence into the hands of AI agents and the people who use them — hasn't been. HederaToolbox closes that gap. 23 tools. Pay per call. No registration. Send HBAR and you're in.
Hedera's infrastructure is world-class. But between raw on-chain data and someone actually using it — a compliance officer, an AI agent, a developer — there's a gap nobody built a bridge for. Until now.
Hedera exposes REST and gRPC APIs, not MCP tools. Agents can't call them directly without a translation layer.
Mirror node responses are verbose and unstructured. They require domain knowledge to interpret correctly before an agent can act on them.
API key forms, account setup, and monthly invoices slow down agent deployments — especially autonomous ones that need to self-onboard.
Writing an auditable on-chain record and verifying it later requires HCS expertise most engineering teams don't have on hand.
From a terminal agent running autonomously on Hedera mainnet, to a compliance workflow in Claude Desktop, to a casual conversation in Claude.ai — all powered by the same 20 tools.
A single business objective — onboard a counterparty, produce a compliance clearance — pasted into Claude Desktop. No steps specified, no tools named. The agent decides what to call, in what order, narrates its reasoning, flags risks it wasn't told to look for, and delivers a board-ready summary with a tamper-proof record written to Hedera mainnet.
Add HederaToolbox as a connector in Claude.ai, send HBAR to the platform wallet, and start asking questions. No terminal. No configuration. Live on-chain intelligence surfaced conversationally — token prices, risk signals, supply key flags, governance proposals.
A 7-step autonomous compliance workflow running as a standalone Node.js script in a terminal. No Claude Desktop, no human in the loop — just an agent calling the MCP server directly over the network, exactly how a production deployment would run.
Every tool returns a clean, structured response. No mirror node parsing. No schema work. Your agent calls the tool and gets back something it can act on.
HederaToolbox speaks MCP over HTTP/SSE — the open standard every major AI platform now supports. Connect once, call any of the 20 tools from wherever your agents live.
Add as a custom connector in Settings. Paste the endpoint URL — no package install, no config file. Works on claude.ai web, mobile, and desktop app.
Install the npm package globally, then add the config block to claude_desktop_config.json. Restart Claude Desktop and the tools appear automatically.
Settings → Connectors → Add. Set transport to HTTP/SSE, paste the endpoint. Manus auto-discovers all 20 tools and their schemas instantly.
Add as an MCP server in Cursor settings. Cursor supports remote MCP servers natively — no local install required.
If your agent supports MCP over HTTP/SSE, it works. Point it at the endpoint — tools are auto-discovered from the server manifest.
No UI needed. Use the npm package directly in Node.js — Railway, cron, Docker. The one-file agent template in the GitHub repo gets you started in minutes.
View example scripts on GitHub ↗Works with Claude Desktop, Cursor, OpenClaw, or any MCP-compatible agent. No server setup required.
Claude.ai: Settings → Connectors → Add → paste the endpoint URL. Claude Desktop: run npm install -g @hederatoolbox/platform first, then add the config block below to claude_desktop_config.json.
Send any amount from your Hedera account. No form, no registration. The deposit watcher checks every 10 seconds.
Pass it as api_key in any tool call. Balance deducted per call. Top up any time.
https://api.hederatoolbox.com/mcpnpm install -g @hederatoolbox/platform{
"mcpServers": {
"hederatoolbox": {
"command": "npx",
"args": ["-y", "@hederatoolbox/platform"]
}
}
}clawhub install mountainmystic/hederatoolboxServer Name: hederatoolbox
Transport: HTTP/SSE ← not STDIO
Connector URL: https://api.hederatoolbox.com/mcpIf your agent needs to self-discover the platform without parsing this page, use the machine-readable file at hederatoolbox.com/llms.txt
You authenticate with your Hedera account ID. That's a public identifier, like a username. Your private key stays on your machine, in your wallet, under your control. HederaToolbox has no way to access it, store it, or ask for it.
No tiers, no seats, no monthly invoice. Balance is deducted per tool call. Top up any time by sending HBAR to the platform wallet.
| Tool | Module | Description | Cost |
|---|---|---|---|
| account_info | Platform | Wallet address, pricing, your current balance | Free |
| get_terms | Platform | Machine-readable terms of service | Free |
| confirm_terms | Platform | Record consent before paid calls | Free |
| hcs_monitor | HCS | Topic status and recent activity | 0.10 ℏ |
| hcs_query | HCS | Natural language search over HCS messages | 0.10 ℏ |
| hcs_understand | HCS | Anomaly detection, trend analysis, risk assessment | 1.00 ℏ |
| hcs_write_record | Compliance | Write tamper-evident record to any open HCS topic — agent history, continuity, audit logs | 0.10 ℏ |
| hcs_verify_record | Compliance | Verify a record has not been tampered with | 1.00 ℏ |
| hcs_audit_trail | Compliance | Full chronological audit trail for an entity | 2.00 ℏ |
| identity_resolve | Identity | Account profile, token holdings, transaction history | 0.20 ℏ |
| identity_verify_kyc | Identity | KYC grant status for a Hedera account | 0.50 ℏ |
| identity_check_sanctions | Identity | On-chain risk signals and counterparty screening | 1.00 ℏ |
| token_price | Token | Live price, market cap, 24h trading volume | 0.10 ℏ |
| token_monitor | Token | Whale movements and unusual trading patterns | 0.20 ℏ |
| token_analyze | Token | Holder distribution, transfer velocity, risk score | 0.60 ℏ |
| governance_monitor | Governance | Open proposals, deadlines, current vote tallies | 0.20 ℏ |
| governance_analyze | Governance | Voter sentiment, participation rate, outcome prediction | 1.00 ℏ |
| contract_read | Contract | Contract info, bytecode size, storage details | 0.20 ℏ |
| contract_call | Contract | Read-only function call, no gas cost | 1.00 ℏ |
| contract_analyze | Contract | Activity patterns, caller distribution, risk assessment | 1.50 ℏ |
| fixatum_register | Fixatum | Register agent DID — permanent W3C identity on Hedera HCS | 100 ℏ |
| fixatum_score | Fixatum | Live KYA trust score and component breakdown | Free |
| fixatum_status | Fixatum | Registration status, score, and provenance binding | Free |
From enterprise compliance teams to developers building AI agents to anyone curious about what's happening on Hedera — the same 20 tools, the same pay-per-call model, the same live on-chain data.
Council members and regulated businesses. You need monitoring, audit tooling, and compliance records without custom integration work.
Building AI agents that need Hedera data. You don't want to maintain a mirror node integration. Call a tool, get a usable answer.
You hold HBAR or just want to know what's happening on Hedera. Add the Claude.ai connector and ask in plain English.
The HederaToolbox Telegram bot can answer questions about the platform, walk you through getting connected, and flag anything that needs the team's attention directly.
Ask in plain English. The bot knows the full tool list, pricing, connection config, and authentication mechanics. It answers immediately and escalates anything it can't handle to the team.
If you're evaluating HederaToolbox for a team or enterprise deployment — volume pricing, dedicated support, custom integration — the bot will take your details and route them directly to the team.
HederaToolbox gives your agent live on-chain data — token intelligence, compliance records, contract analysis, governance signals. But a capable agent also needs a verifiable identity: a permanent DID and a trust score that tells every platform it encounters exactly who it is and how it’s behaved.
Fixatum is the identity layer built on top of HederaToolbox. Register your agent’s DID once. Every verified Toolbox call is logged against it automatically, building a KYA (Know Your Agent) trust score from four components: activity provenance (0–40 pts), account age, on-chain screening, and anomaly signal. The score rewards consistent, diverse activity — and decays if an agent goes dormant.
Get your agent a DID at fixatum.com →New to agent development? The hedera-agent-starter scaffold gives you Toolbox integration, Fixatum registration, and a working heartbeat loop in one fork.
20 tools. Live on-chain intelligence. Pay per call. Your agent knows what’s happening on Hedera.
W3C DID anchored to Hedera HCS. KYA trust score built from verified Toolbox activity. Every platform knows who your agent is.
3–5 second finality. Fees at fractions of a cent. Enterprise governance. The only chain built for agent-scale throughput.