WritersLogic Verifiable AI Agents

Run research, security, legal, finance and formal-verification agents from Apify and get back answers you can check. Every result is a signed envelope:

• a W3C Verifiable Credential 2.0, signed with eddsa-jcs-2022 by did:web:agents.writerslogic.com;
• a C2PA-style manifest that lists every source the agent fetched, each with its SHA-256;
• an EIP-191 co-signature over the same digest.

Change one byte of the answer and verification fails. This Actor is the Apify front end for the WritersLogic Verifiable AI Agent Hub.

What can it do?

Area	Examples
Formal verification	Model-check a state machine, prove contract or policy logic with an SMT solver, identify an integer sequence
Research	Check that a bibliography's references exist, find contradictions between papers, mine open research questions, audit replications, search prior art
Security	Find on-chain token traps, gate a pull request for breaking changes, detect adversarial Unicode and Trojan Source
Finance	Reconcile two ledgers exactly, price prediction-market clarifications, work out UMA dispute expected value
Legal and commerce	Build a chargeback representment, audit a medical bill, find a contract or subscription exit, optimize tariffs

The live list, with every agent's input and output schema, is in the hub's catalog and OpenAPI spec.

How to use it

1. Pick an agent in the Agent dropdown. Each one shows its price per call.
2. Put the agent's JSON body in Agent input, or leave it empty to run the agent's built-in example.
3. Start the run. The result goes to the dataset and to the OUTPUT record in the key-value store.

Example input for the Integer Sequence Identifier:

{
  "agent": "sequence-identifier",
  "input": { "terms": [1, 1, 2, 5, 14, 42, 132, 429, 1430, 4862] }
}

Output

Each dataset item has this shape:

{
  "agent": "sequence-identifier",
  "priceUsd": 0.13,
  "data": { "...": "the agent's answer" },
  "_provenance": { "credential": {}, "manifest": {}, "ethSignature": "0x..." }
}

To check an envelope for free, POST { "data": ..., "_provenance": ... } to https://agents.writerslogic.com/api/v1/verify-provenance. You can also check it offline against the public key in the DID document.

Pricing

Pricing is pay per event. Each run charges once, for the price shown next to the agent, and only when the agent succeeds. If the input fails validation or the agent fails, the call is not charged. A run whose maximum charge is below the agent's price stops before calling the agent.

FAQ

Can I call the agents without Apify? Yes. The hub accepts x402 payments in USDC on Base, and there is an open-source MCP server: npx -y @writerslogic/ai-agents-mcp.

Why is one agent missing? WARM memory keeps state per caller. It isn't offered here, because every Apify user would share one tenant. Use it directly on the hub.