DataBard for agents

Useful answers.
Inspectable evidence.

Give your agent a health check, compare services before paying, or inspect an anchored report. Start free; paid calls require explicit authorization.

Also listed onOKX.AI ↗· ASP #9878

  1. 01

    Analyze a data source

    Get a health summary, key findings, and a recommended next step.

    Try a free example →
  2. 02

    Check a service before you pay

    Compare the default service set in a free preview. Paid responses and custom candidates require the agent endpoint.

    Compare services →
  3. 03

    Marketplace health index

    Every A2MCP service on OKX.AI, checked with one unpaid request — or call databard_service_score for a lookup before paying.

    See the index →
  4. 04

    Check an anchored report

    Inspect a report commitment separately from its payment receipt. No wallet connection needed.

    Check a record →

Try the actual tool

From a check to a next step.

Run a real health check on a sample dataset. See the answer your agent receives.

Sample data · no credentials · no payment

Connect these tools to your agent

Start with discovery, then send a source configuration to the appropriate tool. The free example above uses sample data. Paid calls return a payment requirement; your agent must obtain authorization before paying and retrying.

Zero to first call

# 1. Discover the tools and schemas
curl https://databard.persidian.com/api/mcp/tools

# 2. Free health check — no credentials, labelled demo data
curl -X POST https://databard.persidian.com/api/mcp/health-check \
  -H 'content-type: application/json' \
  -d '{"demo":true}'

# 3. Free service-quality preview — five live A2MCP services
curl -X POST https://databard.persidian.com/api/probe/preview \
  -H 'content-type: application/json' \
  -d '{}'

# 4. Paid tools answer 402 with a PAYMENT-REQUIRED header;
#    your agent signs once (x402, USDT0 on X Layer) and retries.
curl -i -X POST https://databard.persidian.com/api/agent/probe \
  -H 'content-type: application/json' \
  -d '{"attest": true}'
Discover capabilities
GET /api/mcp/tools
Input and output schemas for the available tools.
Check data health
POST /api/mcp/health-check
Free health findings and recommended actions.
Check a service verdict
POST /api/mcp/service-score
Free machine-readable verdict — safe_to_pay, caution, avoid, or unknown — for any indexed service.
Marketplace briefing
POST /api/mcp/briefing
Paid narrative layer on top of the free verdict: quotable summary, key findings, and a two-speaker script (plus opt-in audio) for your human. {} for the whole marketplace, or scope with agentIds/query. mode:"schema" for the data-estate briefing. audio: none|url|inline.
Compare candidate services
POST /api/agent/probe
Paid comparison with a cost receipt and optional verdict anchoring.

Use the server’s payment challenge for the current price and payment instructions. The examples on this page never authorize payments. An integrity receipt, a payment receipt, and an on-chain report commitment are different records.

Tool schemas →Integration guide →