Home / Directory / AI SaaS tooling / Document intelligence / IDP / Unstract
Unstract
Unstract is an open-source document platform from Zipstack, in Los Altos, California, that turns PDFs, scans, and other files into structured data with large language models, and pairs a prompt-based Document Studio with LLMWhisperer, a parser that preserves page layout.
Engineering and operations teams extracting data from invoices, bank statements, forms, and similar files, who want open-source code, a prompt-based workflow, and the option to self-host.
Teams that want a finished, per-document-type product with no prompt writing, and buyers who need the Flow agent product priced before a sales call.
Verdict
Unstract is worth a look when a team wants control over how documents become data. Its Document Studio code is open source under AGPL-3.0, its cloud plans post monthly prices and per-page overage rates, and LLMWhisperer posts per-page parsing prices with 100 free pages a day. Heavy users should price the overage rates against their page volume.
Score Breakdown
These scores cover how Unstract handles extraction workflows, parsing of messy layouts, deployment choice, and human review and integrations, and then how it does as a company on innovation, communication, pricing, and contract terms.
Who Uses Unstract, Where It Fits on Price, and Its Competitors
Unstract is built by Zipstack and has three parts. Document Studio is where teams write prompts that extract fields from documents, review results with a human in the loop, and publish the workflow as an API, an ETL job, or an MCP server. LLMWhisperer is a parser that converts scans and PDFs into text that keeps their layout so a model can read tables and forms. Flow, the newest product, builds agentic workflows. Unstract's homepage says it has more than 600 paying customers and processes more than 10 million pages a month, both company figures, and its GitHub repository shows about 7,300 stars.
Document Studio costs $499 a month for 5,000 pages and $2,249 a month for 25,000 pages on monthly billing, with an Enterprise plan by quote. Its closest competitors are Extend, Reducto, Hyperscience, and Indico Data, which score higher in our reviews, and Affinda, which scores lower, and Docsumo is another document platform we reviewed that quotes prices by volume.
Pricing
We checked Unstract's pricing page and recorded the Document Studio plans, the Enterprise plan, and the LLMWhisperer parsing rates.
| Plan | Price | What stands out |
|---|---|---|
| Document Studio Starter | $499 a month | 5,000 pages a month, then $0.10 per page, with a 14-day free trial and no credit card needed. |
| Document Studio Growth | $2,249 a month | 25,000 pages a month, then $0.09 per page, on monthly billing. |
| Document Studio Enterprise | Custom | Cloud or self-hosted, with SOC 2, GDPR, HIPAA, and ISO 27001 compliance and priority, dedicated support. |
| LLMWhisperer | $1 to $15 per 1,000 pages | Native text pages cost $1 per 1,000, low-cost scans $5, high-quality scans $10, and form and table parsing $15. The page lists 100 free pages a day. |
| Flow | Talk to us | The pricing page lists Flow without a posted price. |
The Field at a Glance
How Unstract compares with other document intelligence vendors we reviewed, by Overall Score and relative typical engagement cost.
Unstract scores 7.6 in document intelligence, under Extend (7.9), Reducto (7.9), Hyperscience (7.8), and Indico Data (7.7), and ahead of Docsumo (7.3), Affinda (6.6), and Nanonets (6.5). Docsumo publishes no list price, so it appears at the far right as No published price.
Compared with …
- Unstract vs Docsumo 7.6/7.3
How Well Unstract Fits Each Document Extraction Use Case
We rated Unstract against the situations teams bring to a document extraction platform, drawing on its homepage, pricing page, and GitHub repository and public developer discussion.
| Use case | Fit | Notes |
|---|---|---|
| Extracting fields from invoices and bank statements | Strong | Document Studio extracts fields with prompts, and the pricing page points to an API Hub with ready-made APIs for bank statements, tables, and invoices. |
| Scanned and messy PDFs | Strong | LLMWhisperer keeps layout and can fall back to OCR for scans, which a Hacker News commenter praised in 2024. |
| Self-hosting for data control | Strong | The code is open source under AGPL-3.0, and the Enterprise plan runs in the cloud or self-hosted. |
| Human review of uncertain results | Mixed | Human-in-the-loop review is part of Document Studio, and the review tooling is less prominent than at managed IDP vendors. |
| Teams without engineers | Weak | Prompts, APIs, and workflow setup assume technical staff. |
| High page volumes at a predictable cost | Mixed | Page prices are posted, and overage at $0.09 to $0.10 per page adds up past the plan limits. |
Review Excerpts
These excerpts come from two Hacker News comments from June 2024 by an independent user and from the buyer concern we found on Unstract's pricing page. Paid reviews and pay-for-play sites were excluded.
“LLMWhisperer has some nice tooling where they can fall back to OCR as well forcing text extraction from scanned documents as well as documents that have the text preserved as text.”
“LLMWhisperer is great, seconded.”
“The page prices are posted, but past 5,000 or 25,000 pages the overage adds up, and Flow has no price at all.”
Where customer reviews couldn't be found, our analysts attempted to give a voice to buyer sentiment, drawing on real information from review sites, forums, and product documentation.
Based on: Unstract pricing
Company Facts
- Founders
- Shuveb Hussain, Arun Venkatasamy, Narendran Hariparanthaman, and Rao Cherukuri (executive chairman) (Together Fund)
- Headquarters
- Los Altos, California, with most of the team in India (LinkedIn)
- Funding
- A seed round in November 2022 backed by Together Fund and Lightspeed India Partners, which Crunchbase lists at $5.24 million (Crunchbase)
- Products
- Document Studio, LLMWhisperer, and Flow (Unstract)
- Open source
- AGPL-3.0 code on GitHub (GitHub)
- Website
- unstract.com
Key Leadership
Rao Cherukuri
Co-founder and executive chairman
Listed that way on Together Fund's Zipstack page (Together Fund).
Methodology
We scored Unstract on eight criteria, from extraction workflow to contract fairness, and the Overall Score is their simple average on a 0 to 10 scale. The scores draw on reviews from public review sites and forums, with paid reviews and pay-for-play sites excluded, plus a hands-on look at the product, its pricing, and its documentation and training. Scores can change as we add reviews and product checks.