Dynamic LLM Routing
Auto-selects the optimal model per document region — text, tables, images, mixed content.
Magic
Ingest.Extract.Search.
DocSlurp transforms any document into structured, searchable intelligence. Smart LLM routing, table extraction, image captioning, citations — all through one API.
quickstart.ts
import { createDocSlurpClient } from '@docslurp/sdk';
import { readFile } from 'node:fs/promises';
const client = createDocSlurpClient({
apiUrl: process.env.DOCSLURP_URL!,
apiKey: process.env.DOCSLURP_API_KEY,
});
const { workspace, run } = await client.upload({
files: [{ filename: 'report.pdf',
data: new Blob([await readFile('report.pdf')]) }],
});
await client.waitForRun(run.id);
const answer = await client.chat({
workspaceId: workspace.id,
q: 'What are the key findings?',
});PDF to structured data
in one API call.
From raw PDFs to structured, searchable intelligence — DocSlurp handles the entire pipeline.
Advanced Table Extraction
Variable columns, merged cells, mixed content types — all handled with structural fidelity.
Image Intelligence
Turn document images into searchable context alongside extracted text.
Citations & Page Images
Every search result links back to exact page regions with rendered page image previews.
Smart Region Detection
Identifies headers, footers, sidebars, figures, and body text with spatial precision.
Enrichment Pipeline
Entity extraction, sentiment analysis, topic classification — layered onto every document.
Interactive Word Clouds
Dynamic, filterable word cloud visualizations for rapid corpus understanding.
Subject & Intent Graphs
Explore relationships between topics, entities, and document intent in interactive graphs.
Optimized Processing
Parallel extraction pipelines with intelligent caching for sub-second re-queries.
Focused Retrieval
Search within a workspace and narrow results by document, filename, or tags.
Agent Tools
Authenticated REST tools for workspace search and grounded agent integrations.
SDK, API & CLI
TypeScript SDK, REST API, and CLI — pick your interface, ship in minutes.
The pipeline
From upload to searchable documents, with sensible defaults.
Step 1
Ingest
Upload PDFs, DOCX, images, HTML — any document format.
Format detectionParallel chunkingPage rendering
Step 2
Process
LLM-powered extraction with smart model routing.
Region detectionTable parsingImage captioning
Step 3
Enrich
Entity extraction, classification, and relationship mapping.
NER & sentimentTopic graphsWord clouds
Step 4
Search
Semantic search with citations, page images, and filters.
Vector + keywordCited resultsFaceted filters
Three interfaces. One platform.
SDK for developers, REST for integrations, tools for AI agents — choose your path.
SDK
TypeScript SDK
Upload, wait for processing, search, and ask questions with a small typed client.
import { createDocSlurpClient } from '@docslurp/sdk';
const client = createDocSlurpClient({
apiUrl: process.env.DOCSLURP_URL!,
apiKey: process.env.DOCSLURP_API_KEY,
});
const { results } = await client.search({
workspaceId: 'workspace-id',
q: 'revenue Q4',
});
for (const hit of results) {
console.log(hit.text, hit.citation);
}API
REST API
Upload batches, track processing, and retrieve cited answers through REST.
curl -X POST "$DOCSLURP_URL/v1/ingest/bulk" \
-H "Authorization: Bearer $DOCSLURP_API_KEY" \
-F "[email protected]"
# Returns workspace, run, and documents.
# Check /v1/runs/<run-id>/progress
# before searching your workspace.AGENTS
Agent Tools
Authenticated REST tools let your agents search the same workspace documents.
curl -X POST \
"$DOCSLURP_URL/v1/mcp/tools/workspace.search" \
-H "Authorization: Bearer $DOCSLURP_API_KEY" \
-H "Content-Type: application/json" \
-d '{"workspaceId":"workspace-id",
"q":"revenue Q4"}'
# REST tools for agent integrations.Start building in minutes.
Drop files, auto-create workspaces, and surface grounded search — no configuration required.