Memory your AI agent can prove: pay-per-call storage and a signed spending ledger over x402

Description and use

Tabularium lets agents persist data with a verifiable hash chain and maintain a signed spending ledger, ensuring transparency and auditability of paid services.

Problem

Agents need a way to persist learned data across sessions and to provide verifiable records of spending when they use paid services.

Outcome

Tabularium enables agents to persist data with verifiable proof and to maintain a signed spending ledger, providing transparency and auditability for paid services.

Workflow

• 1. Add Tabularium to the MCP client with the command: claude mcp add tabularium --env TABULARIUM_PRIVATE_KEY=0x... -- npx -y tabularium-mcp
• 2. Ask the agent to remember data or use the load_memory prompt.
• 3. The agent performs write/read calls; each write is appended to a hash chain and signed by the server.
• 4. The MCP server writes a signed receipt into the agent’s workspace via tabularium_pay for each purchase.
• 5. The agent can export its entire history; anyone can verify it offline with a zero‑dependency script.
• 6. The tabularium_ledger totals spend by service and day.
• 7. Spending is capped at $1 per session by default, with a single purchase limit of $0.10.
• 8. The service is paid per call over x402 in USDC on Base; a typical session of 10 reads and 20 writes costs about eight cents.

Evidence

A small paid service with four calls: write, read, renew, export.

Source

Every write is appended to a hash chain, and the server signs the head with an Ed25519 key published in its menu.

Source

The MCP server includes tabularium_pay: it pays any x402 service and writes a signed receipt into the agent's workspace...

Source

From Claude Code or any MCP client: claude mcp add tabularium --env TABULARIUM_PRIVATE_KEY=0x... -- npx -y tabularium-mcp

Source