LlamaParse
Hosted parsing API that turns complex documents into Markdown for AI pipelines.
Operated by LlamaIndex
Where your files are processed
Reported to upload your files
The vendor states that files are uploaded to their servers. We have not independently tested this.
A hosted API: documents are uploaded to LlamaIndex's service and parsed there. The vendor lists "local cloud deployment" among its enterprise features, which indicates a self-hosted option exists for larger customers, but the default and documented path is their service.
What it does
LlamaParse takes a document — a dense financial table, a scanned form, a paper full of equations — and returns clean Markdown that a language model can work with. It is a hosted service reached through an API or one of the vendor's SDKs, with several parsing modes trading cost against accuracy, and it claims support for more than ninety file formats and a hundred languages. The company that runs it also maintains the LlamaIndex framework, so it is the parsing step of a wider set of tools for building applications on documents.
Best for: Teams building retrieval or extraction pipelines who want document parsing as an API call rather than a model to run, and who need hard layouts handled — tables, multi-column pages, handwriting and checkboxes are the four the vendor names. The parsing modes are the practical selling point: the same document can be parsed cheaply or carefully depending on what it is for.
Worth knowing: It is a hosted service, so documents are uploaded, and for anything confidential that is the first thing to settle — the vendor mentions local cloud deployment as an enterprise option, which implies the default is theirs. Pricing is published: a free plan with 10,000 credits that is not a trial, Starter at $50 a month and Pro at $500 a month, with custom Enterprise terms. Credits are the unit rather than pages, and what an action costs depends on the operation and the mode, so a credit count is not a page count. The Enterprise option is documented as private VPC or hybrid cloud, which is not the same as an offline downloadable edition. The accuracy claims are the vendor's own, including a benchmark the vendor publishes, which is worth knowing before treating a comparison as independent. And the product's name is moving: the page is at /llamaparse and the navigation lists LlamaParse, but the page's own copy calls the product Parse — a reader following documentation written against either name should check which one they are reading.
Details, and where each comes from
Vendor statements, our own reading of their pages, and the things nobody has established yet, each answered separately. Where we have not confirmed something it says so rather than guessing, and a row nobody has established is greyed out.
- Versions covered
- API
- Website
- llamaindex.ai
- Requires an account
- Not verified
- Free tier
- Yes
- Watermark on the free tier
- Not verified
- Maximum file size
- Not verified
- Batch processing
- Yes
- Data residency
- Not verified
- Open source
- No
- Self-hostable
- Yes
- Licence
- Commercial, usage-based; free entry point via sign-up
- Redaction
- No redaction feature
- Programmatic access
- HTTP API
- Last availability check
- Responded in 67 ms 16 October 2026
Checked automatically by our own runner, not reported by the vendor. 100% of our last 30 checks got a response.
What the redaction verdict rests on
From the vendor's own documentation, read 22 September 2026. Not tested by us.
“Parse transforms messy documents into AI-ready data at scale”
What that does not establish: Recorded as `none` on absence: the product reads documents and emits structured text, and nothing on the page describes removing content from a file or producing a modified document. What it does not establish: anything about the safety of the output, and there is a specific reason to say so here rather than in passing. A parsing service whose purpose is to extract everything faithfully is the opposite of a redaction tool, and its output — clean Markdown containing the document's contents — is precisely the artefact that must not be published when a document was supposed to be redacted.
What the vendor says
Claims are labelled rather than repeated as fact. Where we have checked something ourselves, it says so and shows when. Checked means we read the vendor's own page and it says what we recorded — not that we tested the behaviour, unless the note says we did.
-
“Parse transforms messy documents into AI-ready data at scale. Turn even the most complex layouts, tables, charts, handwriting, checkboxes, and images into clean markdown”
Checked by us — the source says this Read 1 October 2026
Confirmed. The quoted text appears in the rendered page as read on 2026-10-01. This claim was carried as reported rather than checked until this read.
Standing: Not re-read against the live source
Check it yourself: the source, the SHA-256, and how we checked
- From
- the text a browser renders
- SHA-256
- e84d96481c03c0542679ff33cd60449b7b55291dc5fcc6fa326fa8ee048f851a
How we checked: Read as rendered text in headless Chrome on 2026-10-01 and saved to the evidence archive, so the read can be repeated against the same file. Comparison ignores case, punctuation and hyphenation, which is why "open-source" on the record matches "open source" on the page.
-
“Best-in-class accuracy with support for 90+ formats”
Checked by us — the source says this Read 1 October 2026
Confirmed. The quoted text appears in the rendered page as read on 2026-10-01. This claim was carried as reported rather than checked until this read.
Standing: Not re-read against the live source
Check it yourself: the source, the SHA-256, and how we checked
- From
- the text a browser renders
- SHA-256
- e84d96481c03c0542679ff33cd60449b7b55291dc5fcc6fa326fa8ee048f851a
How we checked: Read as rendered text in headless Chrome on 2026-10-01 and saved to the evidence archive, so the read can be repeated against the same file. Comparison ignores case, punctuation and hyphenation, which is why "open-source" on the record matches "open source" on the page.
Each claim is quoted from the vendor's own pages. This section describes what we have and have not checked — it is not a judgement about anyone.
Sponsorship is distribution, not an endorsement or a promise of search rankings, AI citations, traffic, or sales. This listing is a free entry, compiled from published sources.