PRODUCT / API & INTEGRATIONS

Every Feature Available as an API Call

Upload filings, poll extraction jobs, fetch structured results, subscribe to docket alerts, and diff two plans. All through documented REST endpoints with interactive Swagger docs. Build TrancheLab into your workflow, not around it.

View API Docs Book a Demo

Terminal showing sequential API calls: upload PDF, poll status, fetch structured results

7 endpoint groups

Extract, jobs, results, alerts, docket, diff, export

2 auth methods

API key + OAuth

< 10 min

Typical extraction time

Swagger UI

Interactive docs at /docs

Restructuring Data Trapped in PDFs Is Not Programmable

The data you need lives in 300-page filings on PACER. You cannot query it, pipe it into a model, or build automation around it. Until you extract it, it is not data. It is a document.

Problem: Manual, Unstructured Workflows

Download a PDF from PACER
Open it in a reader, ctrl+F for tranche names
Copy numbers into a spreadsheet by hand
No way to programmatically compare two plan versions
No way to trigger extraction when a new filing drops
Every analyst on the team repeats the same work

Solution: Programmable Extraction Pipeline

Upload a PDF via API, get structured JSON back
Poll job status while extraction runs
Fetch results by job ID in JSON or CSV
Diff two extractions to see what changed between amendments
Subscribe to a case and get webhook notifications on new filings
One extraction, shared across the team

How TrancheLab's API Works

STEP 1:

Upload and extract

POST a PDF to the extract endpoint. TrancheLab returns a job ID immediately. The extraction pipeline runs asynchronously: parsing, section classification, tranche extraction, confidence scoring, deduplication. Poll the job status endpoint or wait for completion.

Docs: Extract Endpoint

POST /api/extract request and response with job ID

STEP 2:

Poll job status

Check the progress of any extraction job. The response includes the current stage (parsing, classifying, extracting, scoring, deduplicating), pages processed, and estimated time remaining. Use this to build progress indicators in your own UI or to gate downstream automation.

Docs: Job Status Endpoint

GET job status response with stage progress visualization

STEP 3:

Fetch structured results

When extraction completes, fetch the full tranche table as JSON or CSV. Every tranche includes name, outstanding balance, interest rate, maturity, seniority ranking, recovery estimate, confidence score, and the raw source excerpt that produced each value.

Docs: Results Endpoint

GET results endpoint returning structured JSON with tranches and confidence scores

STEP 4:

Diff two extractions

Pass two job IDs to the diff endpoint. TrancheLab matches tranches across both extractions and returns a structured comparison: values that changed, tranches that were added or removed, and confidence score deltas. Use this to track what moved between plan amendments without reading both filings.

Docs: Diff Endpoint

POST /api/diff request showing changed values, added tranches, and unchanged count

STEP 5:

Subscribe to docket alerts

Register a subscription for any Chapter 11 case. TrancheLab watches CourtListener for new docket entries and delivers notifications via email, webhook, or both. When a new filing appears, you decide whether to extract it by calling the upload endpoint.

Docs: Alert Subscriptions

RESTful API

Every feature in the UI is available as an API endpoint. Upload filings, poll jobs, fetch results, run diffs, manage alert subscriptions. Standard REST conventions, JSON responses, predictable error codes.

Docs: API Reference

Dual Authentication

Static API keys for simple integrations and scripts. OAuth token-based auth for production applications and multi-user setups. Both methods work across all endpoints.

Docs: Authentication

Interactive Swagger Docs

Full OpenAPI documentation at /docs with try-it-now functionality. Test every endpoint directly in the browser. See request schemas, response shapes, and error codes without writing any code.

Docs: /docs

What you can build

Real workflows running on TrancheLab's API today.

Automation flow from CourtListener alert to webhook to extract to database

Automated filing ingestion

Subscribe to cases via the alert API. When a new disclosure statement or amended plan is filed, your webhook fires. Your script downloads the PDF, uploads it to the extract endpoint, and pipes the structured results into your database. By the time you open your laptop, the cap table is already updated.

Daily diff summary email showing changes across multiple cases

Portfolio-wide diff monitoring

Track five active Chapter 11 cases. When any of them files an amended plan, extract both versions and run the diff endpoint. Flag changes in recovery estimates or tranche definitions and surface them in a daily summary email. One script, five cases, zero manual checking.

Python code snippet fetching tranche data and feeding into a model

Feed extraction results into models

Pull structured tranche data from the results endpoint and feed it into your own recovery model, comp analysis, or pricing tool. Every field comes with a confidence score, so your model can weight high-confidence values differently from flagged ones.

One upload shared across team via job ID

Team-wide shared extractions

One analyst uploads a filing. The extraction runs once. Every other analyst on the team fetches the same results by job ID. No duplicate work. No conflicting spreadsheets. One source of truth with provenance on every value.

API endpoints

Method	Endpoint	Description
POST	/api/extract	Upload a PDF and start extraction
GET	/api/jobs/{id}/status	Poll extraction job status
GET	/api/jobs/{id}/results	Fetch structured tranche results
GET	/api/jobs/{id}/results?format=csv	Download results as CSV
POST	/api/diff	Compare two extraction jobs
POST	/api/alerts/subscribe	Subscribe to case docket alerts
DELETE	/api/alerts/{id}	Unsubscribe from a case
GET	/api/docket/{case_number}	Fetch docket entries for a case

FAQ

JSON and CSV. Both include all extracted fields, confidence scores, and source excerpt references. Specify format via query parameter on the results endpoint.

Not currently enforced beyond reasonable use. This will change as the product moves toward production.

No. The same API key works for both. OAuth tokens also work across both interfaces.

Not yet. The API is standard REST with JSON request/response bodies. An SDK is on the roadmap.

You can upload multiple PDFs by calling the extract endpoint multiple times. Each returns its own job ID. There is no dedicated batch endpoint yet, but you can parallelize uploads on your end.

At /docs on the API server. Full Swagger/OpenAPI documentation with try-it-now on every endpoint.

Build on Structured Bankruptcy Data

Every feature in TrancheLab is an API call. Upload, extract, diff, alert, export. Interactive docs at /docs.

Book a Demo