Indexed 001
Newsletter Intelligence
scribem8.com
Curated newsletter and editorial corpus — semantic search across writers, topics and trends.
12.8M
chunks
1536
dims
cosine
metric
We vectorize high-signal corpora — filings, documentation, directories — and ship them as precision-built indexes. Download a self-hosted HNSW index, or query our hosted vector store over a single API.
Each dataset is cleaned, chunked, embedded and indexed. Plug it into your RAG stack the moment you get access — no pipeline to build.
Curated newsletter and editorial corpus — semantic search across writers, topics and trends.
Original equipment manuals, spec sheets and service docs — vectorized for technical retrieval at scale.
A live registry of production AI agents — searchable by capability, interface and deployment surface.
Source data is cleaned, deduplicated and chunked, then embedded with domain-tuned models into dense vectors.
Vectors are assembled into a hierarchical navigable small-world (HNSW) graph, tuned for recall and latency.
Hit it from your stack — self-hosted or via our API. Approximate nearest-neighbor search returns in single-digit milliseconds.
No scraping, cleaning, chunking, embedding or index-tuning. We ship the finished artifact, versioned and ready.
HNSW graphs tuned per dataset for high recall at single-digit-millisecond search — at billions of vectors.
Self-host the index on MongoDB, pgvector or your engine of choice — or let us host it. Same data, your call.
Datasets are re-embedded and re-indexed on a schedule. Pull a new version or stream deltas — never stale.
Download a ready-built index plus backing-DB bundle. Drop it next to your app and search locally — your data never leaves your perimeter.
Skip the infra entirely. Query our managed vector store over a single authenticated endpoint and scale on demand.
Every dataset carries its lineage. Every query is governed. Self-host for full data residency, or run on isolated, encrypted infrastructure.
Self-hosted indexes never leave your environment. Hosted runs in isolated, region-pinned tenancy.
Source, embedding model, version and checksum are attached to every dataset and every result.
Per-dataset API keys, SSO, role-based access and full query logging for compliance.
Tell us what you're building. We'll set you up with a dataset, an index, or an API key — and a short demo to get you querying fast.