Neruva vs Pinecone Nexus

Pinecone announced it. Neruva shipped it.

Pinecone's "knowledge infrastructure for agents" (Nexus + KnowQL) launched May 2026 as Early Access only -- gated to financial services, healthcare, legal, and enterprise SaaS sectors via an application form. Your pcsk_... key doesn't unlock it.

Neruva ships the substrate today. Open API, real benchmarks, sign up in 60 seconds, $5 free credit, no application form.

At a glance

Same frame. Different substrate.

CapabilityPinecone NexusNeruva
AvailabilityEarly Access only -- application + sector qualificationGA today, sign up in 60 sec
Typed records (kind / tags / ts / meta)Marketplace apps + per-field typed outputsOpen Records substrate -- records_ingest / records_query
Per-field citations + confidenceYes (in compiled artifacts)Yes (KG .confidence, analogy .cosine, causal .probability)
Knowledge graph with calibrated confidenceImplied via "artifacts" -- no public KG primitiveSharded HD KG, /v1/hd/kg/*, 100% accuracy at 200 facts
Pearl do-operator (P(Y|do(X))) Not offeredTwo distinct endpoints -- observation vs intervention arithmetically separated
HD analogy (A:B::C:?)Not offered/v1/hd/analogy -- parallelogram completion in HD space
Portable single-file exportHosted-only (BYOC enterprise = self-host stack, not portable file).neruva container -- one zip per namespace, atomic, point-in-time
Deterministic-from-seed replayVersioned artifacts + audit traceBit-identical reruns of the same query (verified 20/20 on /benchmarks)
Open agent-memory protocolKnowQL -- vendor-declared standardTyped-record schema published in docs; any harness can adopt
Auto-record from Claude CodeNot offeredneruva-record-install -- one command, full session capture, secrets-redacted
Cost reduction claim"Up to 90% token reduction" -- single internal financial-analysis test~3,125x cheaper per recall -- measured on /benchmarks against Opus 4.7 list rates
Pricing transparencyBuilder $20/mo flat. Nexus pricing not disclosed.Wallet-based metered, per-op rates published on /pricing
Distribution / ecosystemLangChain, Box, LlamaIndex, Unstructured, Teradata endorsementsMCP install in Claude Code today; Cursor / Codex / Gemini hooks on roadmap
Four substrate primitives Pinecone doesn't have

The substrate is more than a context compiler.

Causal
Pearl's do-operator over HTTP
P(Y | X=1) is what you observed. P(Y | do(X=1)) is what would happen if you forced it. Same logged worlds, two arithmetically distinct endpoints. The killer feature for principled action selection -- the substrate distinguishes co-occurrence from causation natively. Pinecone's artifacts can't do this -- it's a different math, not a different prompt.
Analogy
A:B::C:? in HD vector space
Parallelogram completion via XOR-style algebra over factored items. The substrate finds D by binding arithmetic, not by prompting a model. Returns candidate plus calibrated cosine and an ambiguity gap you can threshold on. No vector DB exposes analogy as a first-class endpoint.
Portable
.neruva -- one file, your namespace
Atomic, point-in-time consistent zip-of-sections containing manifest + records + reserved KG / SCM / analogy slots. Versioned schema. Forward-compatible. Egress is free. Pinecone is hosted-only. Their BYOC option lets you self-host their stack -- it does not give you a portable file.
Deterministic
Bit-identical reruns from a seed
Substrate operations are deterministic in a seed you control. Audit a year later, the substrate returns the same bytes. Verified 20/20 on the public benchmarks page. Pinecone's artifacts are versioned (audit-trace), not seed-deterministic (bit-identical replay).
Where Pinecone is stronger

The honest part.

We're not pretending. Pinecone has things we don't.

Distribution + ecosystem
LangChain, Box, LlamaIndex, Unstructured, Teradata all lined up behind KnowQL at launch. Press cycle wrote the headlines. We're a one-person team with the better product and zero brand recognition. We're catching up by going hard on MCP-host adoption and open-protocol stance.
Pre-built marketplace
Pinecone shipped 90+ vertical apps (Sales / Insurance / Legal / etc.) at launch. We have a /case-study and a benchmark page. If you need a turnkey vertical solution today, Pinecone is faster off the shelf.
Vertical sales motion
Pinecone targets finance / healthcare / legal explicitly with field sales. We're a self-serve product. If you're a regulated enterprise that buys via RFP, you'll find Pinecone's motion easier.
Coming from Pinecone?

Change one import.

Neruva's memory layer is Pinecone-shape compatible. Your existing upsert / query / delete code works unchanged. Then graduate to typed records, KG, causal, and analogy when you want them.

agent.py
# before
from pinecone import Pinecone

# after
from neruva import Pinecone

Or hit our REST directly with the Pinecone host swap: host="https://api.neruva.io" on the Pinecone client. Same JSON shapes. Egress is free. The full mapping is in /integrations.

The cost claim, side by side

~3,125x vs ~10x. One is reproducible.

Pinecone Nexus
~10x cheaper

"Up to 90% token reduction" -- per Pinecone's announcement.

Source: single internal financial-analysis benchmark per The New Stack. Methodology not public.

Neruva
~3,125x cheaper per recall

records_query at $0.000002 / call vs context-stuffing 5 KB at Opus 4.7 input rates ($0.00625 / turn).

Source: live measurements on /benchmarks. Reproducible with one Python file + your own API key.

Don't wait for the application form.

Sign up at app.neruva.io, get $5 free credit, and ship the substrate into your agent in 60 seconds. No sector qualification, no waitlist, no NDA.