Back to MVS
Migrate to MVS in one script
Copy-pastable Python scripts to move every vector from your current database to MVS. Pick your source, run the script, done.
Choose your source database
Each script handles pagination, metadata mapping, and batch upserts. Install pip install mixpeek plus the source SDK, then run.
migrate_pinecone_to_mvs.pyPython
"""Migrate from Pinecone to MVS."""from pinecone import Pineconefrom mixpeek import Mixpeek# ── Connect to both ──────────────────────────────────────pc = Pinecone(api_key="PINECONE_API_KEY")index = pc.Index("my-index")mvs = Mixpeek(api_key="MIXPEEK_API_KEY")NAMESPACE = "ns_migrated"# ── Paginate through all vectors ─────────────────────────batch, cursor = [], Nonewhile True:resp = index.list(namespace="", limit=100, pagination_token=cursor)ids = resp.get("vectors", resp.get("ids", []))if not ids:break# Fetch full vectors + metadatafetched = index.fetch(ids=[v if isinstance(v, str) else v["id"] for v in ids])docs = []for vid, vec in fetched["vectors"].items():docs.append({"id": vid,"dense_embedding": vec["values"],"metadata": vec.get("metadata", {}),})# Upsert to MVS in batchesmvs.namespaces.documents.upsert(namespace=NAMESPACE,documents=docs,)print(f"Migrated {len(docs)} vectors")cursor = resp.get("pagination", {}).get("next")if not cursor:breakprint("Migration complete.")
Prerequisites
1. Install the Mixpeek SDK
pip install mixpeek2. Get your API key
Sign up at studio.mixpeek.com — first 1M vectors are free.
3. Install the source SDK
pip install pinecone4. Run the script
python migrate_pinecone_to_mvs.pyWhy teams migrate to MVS
50x cheaper at scale
Vectors live on object storage, not RAM. 1B vectors for $1,999/mo — not $25,000+.
Zero vendor lock-in
BYO embeddings, BYO object storage. Your data stays in your cloud account.
Upgrade when ready
Start standalone, flip a switch to get extraction, indexing, and pipeline orchestration.