Sieves is live · free · last verified 26 Oct 2026 · viability monitor
Sieves
Open-source Python library for zero-shot document AI, chaining NLP tasks into validated, schema-conformant output without training data.
Sieves is a good fit when you need structured output from documents and have no labeled data to train on. Its Pipeline + Task + Doc model, plus the Bridge layer over DSPy, GLiNER, Hugging Face, LangChain, and Outlines, lets you swap backends without rewriting logic, and the condition parameter stops you paying to classify 40-character documents. Mind the edges: ingestion needs the optional extra (pip install "sieves[ingestion]"), model choice is on you, and the library is openly positioned for rapid prototyping. If you want a hosted document API or an annotation UI, look elsewhere.
Verdict generatedVerified 4d ago · liveness 66/100 · cite: rightaichoice.com/tools/sieves
- Python developers prototyping document AI pipelines
- Data scientists with scarce or no labeled data
- AI consultants scoping document understanding proofs of concept
- Researchers experimenting with zero-shot NLP backends
- Non-technical users who need a no-code interface
- Teams that require a managed hosted API rather than a library
- High-throughput production workloads needing throughput guarantees
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Sieves if you need a no-code annotation tool or a managed API that hides the model layer — Sieves is a Python library where you pick the backends (DSPy, GLiNER, Hugging Face, LangChain, Outlines) and install the parsing extras yourself.
Parsing needs an extra install: Docling and Marker are optional, so budget time for pip install "sieves[ingestion]" plus their own dependency weight.
Sieves is an open-source Python library, so the cost of entry is your engineering time plus whatever you pay the models behind the Bridge layer (DSPy, GLiNER, Hugging Face, LangChain, Outlines). Compare against annotation platforms hired by the seat and per-page commercial document APIs: those bill for labeling labor or page volume, while Sieves bills only through the inference endpoints you already control. It suits individual developers through mid-size data teams; large orgs should factor in
In short
Sieves — Open-source Python library for zero-shot document AI, chaining NLP tasks into validated, schema-conformant output without training data. Best for Python developers prototyping document AI pipelines, Data scientists with scarce or no labeled data, AI consultants scoping document understanding proofs of concept. Free to use.
What is Sieves?
Sieves — Open-source Python library for zero-shot document AI, chaining NLP tasks into validated, schema-conformant output without training data. Best for Python developers prototyping document AI pipelines, Data scientists with scarce or no labeled data, AI consultants scoping document understanding proofs of concept. Free to use.
Sieves is an open-source Python library for zero-shot document AI with structured generation, built for rapid prototyping of document AI pipelines that produce validated output with no training data required. You define the work as a Pipeline — either verbose (Pipeline([t_ingest, t_chunk, t_cls])) or by chaining tasks with + — and Sieves runs them sequentially over Doc objects. Prebuilt predictive tasks cover classification, information extraction, named entity recognition, relation extraction, sentiment analysis, summarization, translation, PII masking, and question answering. Preprocessing splits into ingestion (document parsing via Docling or Marker) and chunking (via Chonkie or NaiveChunker); ingestion libraries are optional and installed with pip install "sieves[ingestion]". A Bridge layer connects tasks to ModelWrapper backends — DSPy, GLiNER, Hugging Face, LangChain, and Outlines — so you can swap the model behind a task without rewriting pipeline logic. Pipelines support conditional task execution through the condition parameter, evaluated per document with lazy evaluation, so you can skip expensive processing on short or off-spec documents while preserving document order and recording results[task_id] = None for skipped items. You also get caching (use_cache, with dump/load/serialize/deserialize and clear_cache), pipeline evaluation, and usage tracking. Sieves is maintained by Mantis, an AI consultancy working on natural language and speech problems. It fits developers and data scientists who need structured output from documents but lack labeled data, rather than teams wanting a no-code annotation workflow or a managed hosted API.
Key features
- Pipeline orchestration via Pipeline([...]) or + chaining
- Zero-shot classification with custom label sets
- Information extraction into structured output
- Named entity recognition (NER)
- Relation extraction
- Sentiment analysis
- Summarization
- Translation
- PII masking for redaction
- Question answering
- Document parsing via Docling or Marker
- Text chunking via Chonkie or NaiveChunker
- Structured generation with schema validation
- Conditional per-document task execution
- Pipeline caching (use_cache, dump/load/serialize/deserialize)
Our long-form take written Jul 22, 2026 is more than 90 days old, so it is being re-reviewed against current data and is not shown.
Limitations
- Sieves is a Python library aimed at rapid prototyping; there is no no-code interface and no managed hosting.
- Ingestion libraries (Docling, Marker) are optional installs, not bundled — you add them with pip install "sieves[ingestion]".
- Parsing and prediction quality depends on the model wrappers and inference endpoints you configure, so you choose and provision the backends.
- The library is maintained by Mantis, an AI consultancy, and the project itself describes its value as rapid prototyping where structured output is needed but training data is scarce, so treat throughput and operational guarantees for large-scale production as your responsibility, not the library's.
as of 2026-09-28
Researching Sieves? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Is Sieves still active?
Sieves is live · free · last verified 26 Oct 2026 · viability monitor.
Viability score
How well maintained and how widely used is Sieves? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →What do users say about Sieves?
What people actually say about Sieves — is it worth it?
Our community sentiment research written Jul 3, 2026 is more than 90 days old, so it is being re-reviewed against current data and is not shown.
How we verified this
Every listed fact about Sieves— pricing, features, links, status — is re-read from the vendor's own pages by our verification pipeline and each change is logged with a date; the last pass ran on 26 Oct 2026. Pages mix model-assisted synthesis with human review; see how we verify.
Verification history
We have re-verified Sieves 9 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — changed SSL certificate
- — changed description, our verdict, features, who it suits, hidden costs, integrations, limitations, switching in, switching out, who should skip it, pricing tiers, pricing fit, setup time, summary, use cases
- — changed domain expiry
- — changed pricing page
- — changed integrations, pricing fit, features, hidden costs, who should skip it, switching in, setup time, switching out, summary
- — changed SSL certificate
Showing the 6 most recent of 9 verification passes.
Free to cite with attribution — this page re-verifies continuously.
How much does Sieves cost in 2026?
Sieves: Free to use. Pricing last verified 26 Oct 2026.
Plans compared
Every published Sieves tier, its price as the vendor publishes it, who it fits, and what it adds over the tier below. The 12-month cost projection follows.
| Plan | Price | Billing | Who it's for |
|---|---|---|---|
| Open Source | Free | — | Individual developers, data scientists and consultancies prototyping document AI on a zero budget for tooling.Adds: Starting tier — the library itself at no license cost, with the model and infrastructure spend kept separate. |
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Where the pricing makes sense
The company stage and team size where Sieves's pricing actually pencils out — and where peers do it cheaper.
Sieves is an open-source Python library, so the cost of entry is your engineering time plus whatever you pay the models behind the Bridge layer (DSPy, GLiNER, Hugging Face, LangChain, Outlines). Compare against annotation platforms hired by the seat and per-page commercial document APIs: those bill for labeling labor or page volume, while Sieves bills only through the inference endpoints you already control. It suits individual developers through mid-size data teams; large orgs should factor in
Has Sieves' pricing changed?
No plan price for Sieves has moved since we began logging in Aug 2026; the pricing page is re-read on every verification pass.
Who should pick Sieves?
Pick Sieves if you are Python developers prototyping document AI pipelines, Data scientists with scarce or no labeled data, AI consultants scoping document understanding proofs of concept, Researchers experimenting with zero-shot NLP backends.
Look elsewhere if you are Non-technical users who need a no-code interface, Teams that require a managed hosted API rather than a library, High-throughput production workloads needing throughput guarantees.
Skip it if Skip Sieves if you need a no-code annotation tool or a managed API that hides the model layer — Sieves is a Python library where you pick the backends (DSPy, GLiNER, Hugging Face, LangChain, Outlines) and install the parsing extras yourself.
GeneratedReal-world workflow fit
Concrete scenarios for the personas Sieves actually fits — and what changes day-one when you adopt it.
Build a Pipeline of Ingestion (Docling, export_format="markdown"), Chunking (Chonkie) and Classification with a zero-shot label set, then run it over a folder of PDFs.
Outcome: Each Doc carries parsed text, chunks and a validated classification in results[task_id], giving you schema-shaped output on day one.
Add a condition callable such as len(doc.text) > 20 to the expensive classification task so only long documents are processed.
Outcome: Short documents are skipped with results[task_id] = None, order is preserved, and the costly model calls are spent only where they matter.
Swap the ModelWrapper behind a task — Outlines, DSPy, GLiNER, Hugging Face or LangChain — and rerun the same Pipeline with use_cache enabled.
Outcome: You compare backends without rewriting task logic, while the cache and usage tracking keep iterations fast and inspectable.
Use cases
- Extract structured fields from contracts without labeled training data.
- Classify support emails into categories with zero-shot labels.
- Mask PII in free-text documents before sharing.
- Summarize long parsed PDFs into short bullet output.
- Translate documents while preserving chunk structure.
- Answer questions against an internal document set.
- Skip expensive classification on documents that fail a length condition.
Setup time & first value
How long it actually takes to get something useful out of Sieves — broken out by persona, not the marketing-page minute.
A developer comfortable with Python gets a first pipeline running in an afternoon: install sieves, add the ingestion extra if you need Docling or Marker parsing, pick a ModelWrapper, and chain Ingestion -> Chunking -> a predictive task. Most of the elapsed time goes to backend selection and credentials rather than library setup, and conditional execution plus caching keep the iteration loop short
Switching to or from Sieves
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From hand-written glue scripts: replace your parse-chunk-classify functions with a Pipeline of Sieves tasks over Doc objects.
- →From Label Studio: keep the structured-output goal but drop the annotation step — Sieves tasks run zero-shot instead of on labeled data.
- →From a per-page commercial document API: move ingestion and extraction in-house via Docling/Marker and a ModelWrapper you control.
- ↗To a managed document API: export your schema and prompts, then reimplement the same extraction behind the vendor's endpoint.
- ↗To Label Studio or a similar annotation platform: add labeled data, then train a task-specific model and retire the zero-shot pipeline.
- ↗To a custom in-house pipeline: Sieves pipelines can be serialized and dumped, giving you a reference implementation to port from.
Integrations
Resources & guides
Tutorials & Learning
YouTube returned 6 videos for “Sieves”, and we withheld 6: 6 could not be judged, because “Sieves” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Sieves.
Official links
What are the best Sieves alternatives?
The closest alternatives to Sieves we list are Litepali, Parsemypdf and Lido. Ranked by direct product-type match, not category overlap.
| Alternative | Pricing | Starts at | Last verified |
|---|---|---|---|
| Litepali | Free | Free | — |
| Parsemypdf | Free | Free | — |
| Lido | Freemium | Pricing not published | — |
Alternatives to Sieves
View all
Litepali
LitePali is a small open-source Python library for image-based document retrieval built on the ColPali architecture.
Parsemypdf
Parsemypdf: an open-source Python PDF parsing library that routes each page to an AI or rule-based backend through one API
Tools that pair well with Sieves
Common stack mates teams adopt alongside Sieves, with the specific reason each pairing earns its keep.
Litepali
LitePali is a small open-source Python library for image-based document retrieval built on the ColPali architecture.
Parsemypdf
Parsemypdf: an open-source Python PDF parsing library that routes each page to an AI or rule-based backend through one API
Lido
Lido extracts data from any PDF, scan, or handwritten document into Excel, Google Sheets, or CSV with no templates and no training.
Sieves vs its closest rival
Sieves's closest head-to-head is with Screenplayiq — If you're a screenwriter or producer needing financial script forecasts, ScreenplayIQ is the clear choice with its box office prediction and pitch deck tools. For developers building NLP pipelines without training data, Sieves offers unmatched zero-shot flexibility and is free. Choose based on your domain: film industry vs. general document AI. Verdict generated
Head-to-head comparisons
Sieves vs Screenplayiq
If you're a screenwriter or producer needing financial script forecasts, ScreenplayIQ is the clear choice with its box office prediction and pitch deck tools. For developers building NLP pipelines without training data, Sieves offers unmatched zero-shot flexibility and is free. Choose based on your domain: film industry vs. general document AI.
GeneratedSieves vs Geologicai
Choose GeologicAI if you are a mining company needing rapid, multi-sensor core analysis with AI logging to accelerate critical minerals projects by over 400%. Choose Sieves if you are a developer prototyping zero-shot document AI pipelines with structured output and no training data requirement — it's free and open source.
GeneratedSieves vs Versatile
Versatile and Sieves are completely incomparable tools serving different ecosystems. If you are a steel erector needing real-time crane visibility without workflow changes, Versatile is your only choice. If you are a developer building zero-shot document AI pipelines with no training data, Sieves wins with its free, flexible library. There is no functional overlap.
GeneratedFrequently asked questions
Best-of guides
Used Sieves? Help shape our editorial sentiment research.