See the whole conversation.
Every post shows sources and why it's here. No black boxes.
Context graph
AI-powered connections
Full sources
Every claim traced
Reproducible
Open source & verifiable
Built for clarity
Context-first architecture that surfaces signal over noise
Context over noise
AI embeddings and a graph of relationships lift signal over sludge. See what matters.
Source trail for every post
Open a context card to see authors, links, relationships, and ranking factors.
Open & reproducible
Export the setup and reproduce results locally. Code, models, and provenance are public.
What you can do
Cut through spam & astroturf
Filter out noise with evidence-based ranking
Find credible voices
Discover topical expertise that's actually earned
Explore the graph
Navigate the full conversation context
How it works
Three-stage pipeline for context-aware content
Ingest
Public, signature‑attested posts only. Verified at the protocol level.
Understand
AI embeddings capture semantic meaning beyond keywords. Deep context awareness.
Rank & explain
Tunable evidence graph and relevance scores surface better results with a "Why this post" receipt.
Vector search with Couchbase 7.6 FTS vector indexes
GPU graph scoring via RAPIDS cuGraph
Model serving with NVIDIA Triton
Optional TensorRT‑LLM optimizations
Provenance trail: model cards, configs, eval reports
OpenShift + GitOps; cloud‑agnostic
Team
Paul Capestany
Founder & Full‑Stack
ex‑SRE; distributed systems & data
Traun Leyden
Senior Software & Deep Learning Engineer
Status
Target latency: sub‑1s P50 for common queries
Open‑source: GitHub (coming soon)
Join the waitlist
Be among the first to experience context-first social