Mistral OCR to MongoDB Atlas: Chunking for RAG
The short answer: Mistral OCR gives you excellent per-page markdown; MongoDB Atlas gives you vector search folded directly into the aggregation pipeline you already run. Wired together naively — flatten, split, embed, insert — they still produce mediocre RAG, because nothing in between rebuilds the document's hierarchy or respects Atlas's projection rules. The missing link is POMA: PrimeCut().ingest() consumes the raw /v1/ocr JSON, emits hierarchy-preserving chunksets, and vektoria's content-free ingest pattern writes only routing metadata into Atlas while assemble() reconstructs prompt-ready context from a volume at query time.
What each end of the pipeline actually provides
Mistral OCR (/v1/ocr, mistral-ocr-latest) returns pages[], each with an index and a markdown string — headings, lists, and tables recovered within each page, image bytes inline when you set include_image_base64. What it doesn't provide: cross-page hierarchy (a heading on page 41 has no machine link to the chapter that opened on page 3) or retrieval units. Details: The Optimal Chunker for Mistral OCR.
MongoDB Atlas provides a $vectorSearch aggregation stage on an Atlas-managed vector index, compound hybrid search via Atlas Search, and ordinary document fields for $match filtering — all inside the aggregation pipeline you already operate. What it doesn't provide: any opinion about what a document should hold, or which fields survive to the output. It ranks whatever you embedded and returns whatever you $project, including nothing at all if you forget. Details: MongoDB Atlas Chunking Strategy for RAG.
The naive wiring, and where it breaks
The common recipe — "\n".join(p["markdown"] for p in pages) → RecursiveCharacterTextSplitter → embed → col.insert_one(...) per chunk → a bare $vectorSearch query — breaks in a pair-specific way:
- Mistral's page indices vanish at the join, so no
$matchpredicate can scope by page, and per-document filtering is all that's left of Atlas's field-level flexibility. - Overlap inflates the collection with near-duplicate documents, each embedded and indexed separately — more vector-index entries to rank through for no retrieval benefit, the same waste any redundant insert would cause.
- The aggregation pipeline drops the exact fields
assemble()needs. A$vectorSearchstage on its own returns Atlas's default projection, not your custommetadatafield — without an explicit{"$project": {"_id": 1, "metadata": 1, "score": {"$meta": "vectorSearchScore"}}}stage immediately after it,file_idandchunkset_indexnever make it into the result, andassemble()returns an empty context list from hits that scored just fine.
The pipeline, end to end
pip install mistralai pymongoimport os
from mistralai import Mistral
from pymongo import MongoClient
from poma import PrimeCut
from poma.vektoria import assemble, records_from_archive, Volume
from poma.embeddings import get_embedder
# 1. Mistral OCR — your existing call, unchanged. Save the raw result JSON.
mistral = Mistral(api_key=os.environ["MISTRAL_API_KEY"])
ocr = mistral.ocr.process(
model="mistral-ocr-latest",
document={"type": "document_url", "document_url": "https://example.com/contract.pdf"},
include_image_base64=True,
)
with open("contract.mistral-ocr.json", "w") as f:
f.write(ocr.model_dump_json())
# 2. The missing link — raw result JSON in, a portable .poma archive out.
client = PrimeCut() # reads POMA_API_KEY
client.ingest("contract.mistral-ocr.json", download_dir="archives", filename="contract.poma")
embedder = get_embedder("local:BAAI/bge-small-en-v1.5")
vol = Volume("s3://your-bucket/poma")
mongo = MongoClient(os.environ["MONGODB_ATLAS_URI"])
col = mongo["poma"]["chunksets"]
# Atlas vector index (numDimensions, similarity) is created once via the Atlas UI/API,
# named "poma_vec" here — not through the pymongo driver itself.
# 3. Content-free ingest — the collection holds only routing metadata; content lives on the volume.
records = records_from_archive("archives/contract.poma") # -> list[Record] (id, text, payload)
for r in records:
vol.write_doc(r.payload["file_id"], {"file_id": r.payload["file_id"],
"chunks": r.payload["chunks"], "text": r.text})
col.insert_one({
"_id": r.id,
"embedding": embedder.embed([r.text])[0],
"metadata": {"file_id": r.payload["file_id"], "chunkset_index": r.payload["chunkset_index"]},
})
# 4. $vectorSearch retrieval — $project the metadata explicitly, then assemble.
qv = embedder.embed(["early termination conditions"])[0]
hits = list(col.aggregate([
{"$vectorSearch": {"index": "poma_vec", "path": "embedding",
"queryVector": qv, "numCandidates": 200, "limit": 10}},
{"$project": {"_id": 1, "metadata": 1, "score": {"$meta": "vectorSearchScore"}}},
]))
context = assemble(hits, volume=vol) # -> [{"file_id", "content"}, ...]POMA validates the Mistral payload shape up front (a corrupted or mislabeled upload 422s immediately), handles all three Mistral image cases, strips running headers/footers, and rebuilds the cross-page heading tree before chunking. Retrieved chunksets share ancestor lineage across a document — assemble() deduplicates that lineage into one prompt-ready cheatsheet, the same discipline that answers our reference legal-document benchmark with 337 tokens instead of 1,542 for a recursive-splitter baseline. Methodology here.
Metadata mapping: POMA fields → MongoDB Atlas primitives
| POMA chunk field | MongoDB Atlas primitive | What it enables |
|---|---|---|
to_embed | embedding field, covered by the Atlas Vector Search index | ANN search via $vectorSearch |
file_id | metadata.file_id, ordinary document field | $match/filter scoping to one document |
page (from Mistral pages[].index) | volume document, not a field | content-free retrieval, cheatsheet assembly |
depth | volume document, not a field | content-free retrieval, cheatsheet assembly |
chunkset_index | metadata.chunkset_index, must be $project-ed | assemble() dedup and ordering |
| chunkset lineage (content) | not stored in the collection — lives on the Volume | content-free documents, smaller cluster footprint |
Frequently asked questions
How do I get Mistral OCR results into MongoDB Atlas for RAG?
Save the raw /v1/ocr JSON, run PrimeCut().ingest() on it, then pull records_from_archive(...), write each chunkset's content to a volume, embed its to_embed text, and insert_one the vector plus a metadata document (file_id, chunkset_index) into an Atlas Vector Search-indexed collection. Retrieve with $vectorSearch and assemble().
Why does POMA's assemble() return nothing from my MongoDB Atlas results after a Mistral OCR ingest?
The aggregation pipeline drops fields you don't project. A $vectorSearch stage must be followed by {"$project": {"_id": 1, "metadata": 1, "score": {"$meta": "vectorSearchScore"}}} — without it, file_id/chunkset_index never reach the result and assemble() gets nothing back.
What MongoDB Atlas fields should Mistral OCR chunksets carry?
An embedding field covered by the Atlas Vector Search index, plus a metadata document holding only file_id and chunkset_index — enough for $match to scope by document alongside the vector stage. page (from Mistral's pages[].index) and depth stay on the volume document with the rest of the chunkset content, not as separate Atlas fields.
Why not just split Mistral OCR's markdown and insert it into MongoDB Atlas directly?
Concatenation discards page indices and heading levels, and splitter overlap inserts near-duplicate documents that inflate the vector index for no retrieval benefit. Atlas then ranks context-free fragments exactly as faithfully as it would rank real chunksets.
Can I combine Mistral OCR content with MongoDB Atlas's hybrid search?
Yes — Atlas Search's compound query type combines BM25-style text matching with $vectorSearch results in one pipeline, useful for OCR'd business documents whose exact tokens (clause numbers, invoice IDs) dense embeddings tend to blur.
Related recipes
Same parser, different store: Mistral OCR → Qdrant · Mistral OCR → Pinecone · Mistral OCR → pgvector
Foundations: The Optimal Chunker for Mistral OCR · MongoDB Atlas Chunking Strategy for RAG · All pipeline recipes · RAG chunking guide · POMA chunksets