Cookbooks - Pulse API

Each cookbook builds one complete workflow, from the first API call to results stored in your own system.

Recipes

Split, then extract

Find each section of a long report, then fill a small schema for each.

Highlight sources

Draw the exact words behind an extracted value on its page.

RAG ingestion

Index a document’s chunks with metadata that cites their pages.

Cloud intake

Parse a whole bucket as one batch and write each result back to it.

Before you start

The recipes are in Python; the TypeScript SDK has the same calls, with camelCase names. Each recipe starts from this client (SDKs covers setup):

import os
from pulse import Pulse

client = Pulse(api_key=os.environ["PULSE_API_KEY"])

import { PulseClient } from "@pulse-software/sdk";

const client = new PulseClient({ apiKey: process.env.PULSE_API_KEY! });

From test to production

Most integrations take this path: Save the settings that work as a saved config, and send its config_id with each pipeline or batch. Then store each result in your own system: a result can expire, and some organizations delete a result the first time it’s read.

Sample documents

Each name opens a hosted test document in the dashboard. To use one in an API call, send https://platform.runpulse.com/api/examples/<id>/pdf as the file_url.

Example	Pages	`id`
Bank Statement	2	`637e5678-30b1-45fa-acc4-877f2d636419`
Rent Roll	10	`18ed11c2-dbce-4bf5-8385-102c55d13480`
10-K Annual Report	12	`0514fc05-8b0a-4a3b-9b9b-18ac89fc04e5`
Complex Table Document	8	`0e36367a-16d2-49c8-99ab-a6fe0df70409`
Research Report	3	`65a31c06-184a-4156-823b-e1686a368b18`
Blackstone Report	6	`b63fca41-ee50-4a65-bd71-64188d5bdbf5`
Attention Is All You Need	15	`3be15d23-d622-4f27-9843-ec2929140eec`
Legal Filing	32	`d4dfc1e2-60ac-4776-a5a8-20b88e68bf9f`

​Recipes