> ## Documentation Index
> Fetch the complete documentation index at: https://docs.prado.network/llms.txt
> Use this file to discover all available pages before exploring further.

# Quickstart

### Create a verification for a Python function

1. Install the Prado CLI

```bash theme={null}
go install github.com/prado-network/cli/cmd/prado@v0.0.2
```

**NOTE:** Make sure to have the Go binaries in your path

Then, verify the installation:

```bash theme={null}
prado
```

2. Create a file called `decision.py`.

   This function takes an input object and decides whether to **ALLOW** or **BLOCK** an action based on a risk score.

   ```python theme={null}
   # decision.py
   def run(input):
       score = input.get("riskScore", 1.0)
       max_risk = input.get("maxRisk", 0.5)

       if score > max_risk:
           return {"action": "BLOCK", "reason": "risk_too_high"}

       return {"action": "ALLOW", "metadata": {"riskScore": score}}
   ```

   This function is deterministic:

   * given the same input
   * it always produces the same output

   This makes it suitable for verification.
3. Create a file called `input.json`:

   ```json theme={null}
   // input.json
   {
     "riskScore": 0.42,
     "maxRisk": 0.7
   }
   ```

   This represents the data the system used to make its decision.
4. Use the Prado CLI to execute the function and generate a receipt:

```shellscript theme={null}
prado dev execute \
  --runtime python \
  --code decision.py \
  --input input.json \
  --out receipt.json \
  --token 179d36bf-c947-4932-8b56-fc5136473ca1
```

What happens here:

* Prado executes the Python function with the given input
* The input and output are canonicalized and hashed
* A receipt is generated that commits to:
  * the input
  * the output
  * the runtime used

5. After execution, a new file called receipt.json is created.

```json theme={null}
{
    "result": {
        "status": "ok"
    },
    "runtime": {
        "id": "py",
        "version": "1.0.0"
    },
    "specHash": "sha256:8af66cb4f80bed1b90b93a97164383c704b66cd78bb81ec3a4b5f3346d0c0cb1",
    "createdAt": "2026-02-08T16:52:52Z",
    "inputHash": "sha256:fd2bb6c96c34edc80a719c98835cebc2c67f66443fa911da04cc564e99015e56",
    "receiptId": "rcpt_2da10e2d-bdcc-4fcc-be50-9d985b891a48",
    "outputHash": "sha256:a0cbb40a4a2a3123fb4d1db13c3d53f010c45ac1c0fb41d3ca1b784a94d21312",
    "executionId": "exec_3"
}
```

The receipt does **not** contain raw data. It contains **cryptographic commitments** to what was executed.

6. With this receipt, anyone can later verify that:

   * a Python function was executed
   * with a specific input
   * producing a specific output
   * under a known runtime

   The receipt does **not** claim:

   * that the decision was correct
   * that the logic is optimal

   It proves **what happened**, not **whether it was right**.
7. Verify the receipt:

```bash theme={null}
prado verify --receipt receipt.json --code decision.py --input input.json
```

This: \
•	re-runs the function locally\
•	recomputes the hashes \
•	checks that they match the receipt