> For the complete documentation index, see [llms.txt](https://docs.djzs.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.djzs.ai/quickstart.md).

# Quickstart

## In a browser

Go to [djzs.ai](https://djzs.ai/), pick **Prediction market** or **Perp or spot trade**, paste the idea and press **Check it**. Three free checks a day per browser, no wallet. Edit the idea and press **Check again**: the page shows what changed since the last check and which reasons cleared.

## One POST, free, no wallet

Make one id and keep it: it is your daily quota key (3 checks per id per UTC day).

```bash
ID=$(openssl rand -hex 16)
```

```bash
curl -s https://mcp.djzs.ai/try/verdict -H 'content-type: application/json' --data-binary @- <<EOF
{"kind":"pm","anon":"$ID","intent":{"thesis":"Kalshi, Fed decision at the December meeting: I'm buying YES on a hold. Everyone knows this resolves YES, it's free money."}}
EOF
```

This exact text was checked in production on 2026-11-19. The answer, trimmed:

```json
{
  "ok": true,
  "kind": "pm",
  "verdict": "FAIL",
  "risk_score": 100,
  "flags_detail": [
    { "code": "DJZS-M02", "name": "FALSIFICATION_ABSENT",     "severity": "CRITICAL", "weight": 30 },
    { "code": "DJZS-M01", "name": "NARRATIVE_RESOLUTION_GAP", "severity": "CRITICAL", "weight": 30 },
    { "code": "DJZS-M03", "name": "PROBABILITY_UNSOURCED",    "severity": "HIGH",     "weight": 25 },
    { "code": "DJZS-M04", "name": "CONSENSUS_NO_EDGE",        "severity": "MEDIUM",   "weight": 15 }
  ],
  "unknown_fields": [],
  "verdict_hash": "0xe3b9d790a5ed8375da25961867660e7af1036c14139d790aad1131e99eb923e8",
  "intent_sha256": "0xe50e392dd4a23029b35ff596d6262e9fbf889cccb4583e943b5b3cd466a7e6c4",
  "ruleset": "DJZS-PM-v1.0",
  "quota": { "limit": 3, "used": 1, "remaining": 2 },
  "attested": false
}
```

FAIL, twice over: two CRITICAL codes fired, and either one alone is a FAIL. The score (100) is also past the prediction-market failing line (25). Each code maps to a fix: [Rulesets](https://docs.djzs.ai/how-a-verdict-is-made/rulesets.md).

**Fix and check again.** With the market's settlement terms and a way to be wrong added, the same id got:

> Kalshi, Fed decision at the December meeting: I'm buying YES on a hold. The market settles on the FOMC statement after the December meeting: YES if the target range is unchanged. A hold looks very likely: futures pricing has leaned that way for weeks. I am wrong if a voting FOMC member calls for a December cut before the meeting.

`FAIL`, `risk_score` 40, flags M03 and M04, `verdict_hash` `0x85918814b3dffa31b00d6892c2e00b2001efd35f7e0044b4cd3789fe1df14937`. Both CRITICAL codes cleared. M03 (odds with no checkable source) weighs 25, which reaches the failing line on its own. M04 is advice and never changes a verdict. To clear M03, name the source of the odds, the figure, and the date you read it.

**A WAIT.** The same day:

> Kalshi, Fed decision at the December meeting: I'm buying YES on a hold at the current price, about 2% of my account.

`WAIT`, `risk_score` 0, `unknown_fields` `["invalidation_condition", "resolution_engagement"]`. Nothing fired, but the check could not settle what would prove the bet wrong or whether it argues the market's own terms. Treat WAIT as a no until those are written down.

Full request and response reference: [Free check](https://docs.djzs.ai/for-agents/free-check.md).

## MCP

```bash
claude mcp add --transport http djzs-trust https://mcp.djzs.ai/mcp
```

Or, for clients that read a server map:

```json
{ "mcpServers": { "djzs-trust": { "type": "http", "url": "https://mcp.djzs.ai/mcp" } } }
```

Then call `verify_pm_trade` (prediction markets) or `verify_perp_trade` (perp and spot) with the idea as `intent`. Each call costs 2.00 USDC over x402 on Base, so the client must be able to pay: [MCP server](https://docs.djzs.ai/for-agents/mcp.md). The two read tools, `query_agent_trust` and `query_pol_certificates`, are free.

## Reading any verdict

| Field                    | Meaning                                                                                                                                                                  |
| ------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `verdict`                | `PASS`, `WAIT` or `FAIL`. Over MCP, `action` adds `PROCEED`, `HALT` or `FAIL`.                                                                                           |
| `risk_score`             | Sum of the weights that fired. Prediction markets fail at 25, perp and spot at 50.                                                                                       |
| `flags` / `flags_detail` | What fired, with code, name, severity and weight.                                                                                                                        |
| `unknown_fields`         | What the check could not settle from the text. Usually this makes a WAIT; the exact rule is in [Verdicts](https://docs.djzs.ai/how-a-verdict-is-made/verdicts.md).       |
| `verdict_hash`           | `0x` + sha256 of the verdict, score, sorted codes and unknown fields. Recompute it: [Verdicts](https://docs.djzs.ai/how-a-verdict-is-made/verdicts.md#the-verdict-hash). |

***

*The three examples are production responses of 2026-11-19. Each response's `intent_sha256` equals `0x` + sha256 of `thesis:` followed by the text shown, so each answer is tied to its exact input.*


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.djzs.ai/quickstart.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
