Report generator

Report generator turns extracted data into a written report. You give it a Markdown template. After Extract finishes, the node fills the template with the extracted values and stores the result as a Markdown file you can download. Saving the workflow stores the template only. No model runs until a document does.

The node

{
  "id": "report_1",
  "type": "report_generator",
  "instructions": "Write amounts with two decimals.",
  "example_report": "# Invoice summary\n\nVendor: {{vendor}}\nTotal: {{total}}"
}

import os

from anyformat.sdk import Client
from anyformat.workflow import Schema

client = Client(api_key=os.environ["ANYFORMAT_API_KEY"])

workflow = (
    client.workflow("Invoice report")
    .parse()
    .extract([Schema.string("vendor", "Vendor name"), Schema.float("total", "Invoice total")])
    .report_generator(
        "Write amounts with two decimals.",
        "# Invoice summary\n\nVendor: {{vendor}}\nTotal: {{total}}",
    )
    .create()
)

import { Anyformat, Schema } from "@anyformat/sdk";

const af = new Anyformat({ apiKey: process.env.ANYFORMAT_API_KEY! });

const workflow = await af
  .workflow("Invoice report")
  .parse()
  .extract([Schema.string("vendor", "Vendor name"), Schema.float("total", "Invoice total")])
  .reportGenerator(
    "Write amounts with two decimals.",
    "# Invoice summary\n\nVendor: {{vendor}}\nTotal: {{total}}",
  )
  .create();

Options

Field	Type	Default	What it does
`example_report`	string	`null`	The Markdown template to fill, up to 64 KiB. Studio and both builders require it to be non-blank.
`instructions`	string	`""`	Extra guidance for filling the template, for example how to format numbers or dates.

The schema the API accepts is generated from the same source: Report generator node schema.

Connects to

ExtractReport generator

Only a single Extract feeds Report generator, and nothing follows it.

Report generator reads the Extract node that feeds it, through exactly one incoming edge. The Python and TypeScript builders add it only to a workflow with a single Extract; send the node in the API JSON to attach it to one branch of a classified workflow.

What it returns

The node fills the reports section of the run results: one entry per split the node generated a report for.

{
  "reports": [
    {
      "node_id": "report_1",
      "split_id": null,
      "status": "succeeded",
      "download_url": "https://storage.anyformat.ai/reports/invoice.md?X-Amz-Signature=…"
    }
  ]
}

status is pending, succeeded or failed. download_url is valid for 15 minutes from the moment the response was built, so re-read the run for a fresh link. In the Python SDK, run.wait().reports holds the same entries.

Get started

Nodes

Workflows

Schema & fields

Documents

Outputs & integrations

Health

Billing

The node

Options

Connects to

What it returns

​The node

​Options

​Connects to

​What it returns

The node

Options

Connects to

What it returns