Sapience Labs

Cultivating collective human wisdom

sa·pi·ence /ˈseɪ·pi·əns/ n.1. wisdom; the trait our species is named for. 2. knowledge grounded in experience, compounded when minds connect.

Every frontier model is trained, then frozen.
We built the part that keeps learning.

A frontier model ships knowing the world up to a cutoff, then stops. Every session you start from zero, re-explaining your project, your constraints, the thing you settled last Tuesday. The field’s own leaders now say it plainly: today’s systems are trained, then frozen, when they should keep learning from experience.

Sapience is continual learning at the system level. It never stops learning from your work, with no model retraining, and nothing leaves your control. What you establish stays with you: structured, cited, and current in every session and every tool you use.

It trains over itself every day, consolidating what you know into sharper knowledge while you are away, the way a brain does at rest. The intelligence compounds instead of resetting.

What you learn stays yours: signed to you, portable across whichever AI you use tomorrow. And when you choose, it connects with other minds, credited, never given away, the first node in a shared memory for humanity.

How it works See it now Ask for a place

Sapience turns what we each know into what we all know. Owned, never taken.

The field's map

The field's map of agent memory.

Every architecture announced this year lives in the first two columns.

In context
retrieved text
similar-looking records
forgets structure, which fact is current, what superseded what
makes one model smarter
In weights
retrained parameters
per-user adapters
forgets everything else, who taught it, when, how to leave the model
makes one model smarter
The missing column
continual learning at the system level
typed, signed knowledge objects
supersession edges, what replaced what at write time
provenance on every fact
portable across every model
owned by its creators, connecting, by choice, across minds
makes what you know compound
01   The tech

What if context weren't limited, and managing it weren't your job?

A system that never stops learning, built like a brain. A frontier model generates from frozen weights: brilliant up to a cutoff, then fixed. Sapience is the part that keeps learning. It is continual learning at the system level, not a bigger model and not a harness around one. It decides what is worth keeping, lets the rest fade, consolidates patterns while you are away, and checks every answer against what it already knows, with no model retraining and nothing leaving your control. Its unit is the Knowledge Object: structured, typed, credited, and linked to others by edges that make discovery possible.

CAPTURE GATE CONSOLIDATE CONNECT CREDIT LEARNING LOOP
Each cycle compounds knowledge, connection, and credit.
i

Prediction-error gating the gate

A prediction-error gate weighs every claim against what the system already knows, so it accumulates signal, not noise.

ii

Consolidation at rest the consolidator

Like a brain asleep, it replays and consolidates episodic memory into durable knowledge, surfacing non-obvious connections. Discovery, not storage.

iii

Belief revision the reviser

When new knowledge contradicts old, it resolves and supersedes rather than holding both. Contradictions are flagged, not buried.

iv

Competitive inhibition the inhibitor

Near-duplicate memories suppress each other, so recall stays sharp as what it knows grows past a single context window.

v

Substitutable cortex the cortex

The reasoning model is interchangeable: Claude, GPT, Gemini, or open. The intelligence lives in the architecture, not in any one vendor's weights.

vi

Attribution built in the registry

Every claim is traceable to its source and cryptographically signable. Provenance is built into the architecture, not an afterthought.

Knowledge object · the unit
“Protein X stabilizes via cooperative binding.”
typefinding
confidence0.92
sourceyour conversation · 2026-05-12
credited toyou
supports cooperative binding as load-sharing supersedes monomer-only stability note bridges alloy stress problem, materials lab

Why knowledge objects, not logs.A knowledge object is the condensed chain of thought behind a result: reasoned over once, at capture, available for recall forever. Flat text stores and session logs save what was said and force every future query to re-derive the thinking. Sapience keeps the thinking itself, which is why what it knows stays small, fast, and reasonable-over at field scale.

A complementary-learning memory architecture, with forty provisional patent filings across it. Performance validation against frontier baselines is ongoing and reported separately.

One mind, whole problem.

Most agent architectures exist to dodge a context limit: split the problem across a swarm, hand each agent a shard, reconcile the lossy handoffs. Every split loses constraints, drifts goals, diverges world models. Sapience holds the whole problem in one reasoning context, ten million tokens of it, past where frontier windows end, so the problem never needs to be split at all. Agents become optional workers, not a structural crutch.

Goals hold.

The whole problem stays in one mind across long work, so objectives do not quietly drift out from under you.

Nothing you established silently disappears.

What you established last month is still there: structured, citable, and current in every session and every tool.

Answers stay grounded.

It reasons from established knowledge and cites it, so it does not invent what it cannot support.

The longer the work, the dumber the AI. Ours holds at 83 to 87 percent across that whole range.

0 50 100% 4K 64K 128K 1M tokens frontier models, reading the full window Sonnet 4.6 · 31% GPT‑5.5 · 17% Gemini 3.1 · 17% DeepSeek · 16% Fable 5 · 64% at 660K Sapience measured: 83–87% from 4K to 1M holds through the window curves illustrative · points measured

An illustration of what we measure. On multi-hop reasoning over accumulated context (matched protocol, three seeds), Sapience holds 83–87% from 4K to one million tokens (84.0% at 1M, n=100) and is measured out to ten million, while frontier models reading the full window fall from 74% at 4K to between 0 and 31% at one million (Claude Sonnet 4.6 across the sweep; GPT‑5.5, Gemini 3.1, DeepSeek V4F, Opus 4.7 at one million).

See a recorded run

And not on one benchmark.

MuSiQue, beyond the windowreal text at 1.5x the reader's window · reproduced across two reader families · official EM, no judge · n=90
38.9%
5.6%
~7x the strongest truncation baseline
Real novels, beyond the windowNoCha-style, public-domain classics · Gemini reader · 3 seeds · n=63
38.6%
20.1%
~1.9x the truncated reader
Single-document QAHelmetHotpotQA · Claude Sonnet reader · n=200 · a loss, disclosed
61.1%
69.6%
-8.5pp; router gate restores parity (validating)
the reader model, on Sapiencethe same model alone

Reader named per row (Claude Sonnet, DeepSeek, Gemini): identical model both arms, only Sapience differs. Numbers locked, three seeds where marked. We publish the losses too; the last row is one of ours. And the obvious alternative fails: recalling only what the question needs holds at scale; dumping everything it knows into context does not.

See a recorded run

Three results, one cause.

Up to 78 points over similarity search

Where answers change over time, traversing supersession edges beats similarity search by up to 78 points. Hand the same model that evidence in the order it was written and the gap disappears. The win is write order, not model capability.

Ten million tokens, zero for the frontier

The same structure carries a frozen model through ten-million-token corpora it otherwise reads at zero, past where every frontier context window ends.

A hundredth the size

The result holds on an open model a hundredth the size of a frontier one. An 8B model lands within a few points of it, at a fraction of the tokens.

See a recorded run, answer by answer

One long-context, generalized reasoning architecture underneath every kind of work: science, engineering, code, analysis, writing.

02   Ready now

What if your AI actually knew your work?

This ships today, not someday. One command, about ten minutes, no API key. It runs behind your Claude Code and Codex sessions today, and researchers are already using it in their daily work. What follows is the live app, not a mockup.

Sapience  ·  your brain
brain.coretx.ai
live · cited to source
Knowledge map
cross-domain link
Research Methods People Bridge Shared by Maya
Ask your brain
What connects my binding-affinity work to the alloy problem?
Answer · your brain + Maya’s (materials science)

Both turn on cooperative binding. Your 2024 note that a protein stabilizes through cooperative binding1 answers Maya, a materials scientist you follow, whose alloy has to hold under stress2. Neither of you went looking; the shared mechanism connected you.3

KO 1Protein stabilizes via cooperative bindingyou · 2024
KO 2Alloy must stay stable under cyclic stressshared by Maya
KO 3Cooperative binding as a load-sharing mechanismthe connection

Today no two people can think together through their AIs. Here is what it looks like when they can: every answer drawn from your own knowledge and the people connected to it, cited to the exact objects it used. Ask a real research brain your own question ›

The Sapience app · in your browser
One brain across every session. Claude Code, Claude, and Codex share the same memory: what one session learns, the next one knows.

Your first week

You stop re-introducing yourself to your AI every morning. It already knows the project, the constraints, the thing you decided last Tuesday.

Your Claude and your GPT finally know the same you. Switch tools mid-thought; nothing resets.

Your plan does more before it runs out, because your AI reads three facts instead of replaying your history.

And the first time a colleague’s knowledge lights up against yours, you feel what the network is for.

03   The product

What if it never forgot, never drifted, and never made things up?

i

A living map

Your work organizes itself into a navigable map, from a single thought to the whole of what you know.

ii

Memory beyond the window

Ten million tokens, years of your work, past where every frontier context window ends and reasoned over at once. Nothing you established is forgotten.

iii

Cited answers, never invented

Every answer is synthesized from your own knowledge and cited to source. When it does not know, it says so.

iv

Discover what you are missing

From any idea, surface the work and people you would not have found by search. For scientists, that reach extends into 250 million papers.

v

One brain, every client

Tell one AI something and every other one knows it: Claude, GPT, Gemini, or open.

vi

Quietly, in the background

It works behind your Claude, GPT, or Codex sessions over MCP, reading what you know and writing back what you learn. Reading the few objects that matter instead of your whole history, answers arrive at roughly 1,800x fewer tokens per query once your accumulated work passes a million tokens (reads scale with the question, not your history; on our benchmark corpus).

vii

Your sessions talk to each other

A session can leave an addressed note for another thread, or a future you. Pinned commitments surface when the next starts, and results route back to the thread that asked, across Claude Code, claude.ai, GPT, and Codex.

viii

Share and import, with credit

Share a project or your whole brain. On import it lights up where a colleague's knowledge meets yours, every claim credited to whoever established it.

04   The network

What if two people could think together through their AIs?

Someone you have never heard of, in a lab or a company you would never think to look in, is working on the other half of your breakthrough. The way it is set up today widens the gap: each scientist several times more productive, the fields they work in measurably less connected (Nature, 2026). Because knowledge stays owned and credited, Sapience matches it across people and organizations no single vendor could connect. In an internal benchmark on discoveries published after the reasoning model’s training cutoff, it ranked the approach that later won first in 11 of 13 problems, on a small internal set whose protocol we share with partners. Watch three strangers solve one problem ›

PHARMA LAB finding: a protein stabilizes via cooperative binding MATERIALS LAB need: an alloy that holds stable under stress SHARED MECHANISM cooperative binding

Memory recalls what one person stored. Sapience matches what everyone knows, by mechanism rather than keyword, and a connection becomes a decision, a collaboration, a breakthrough no one could reach alone.

AI made each of us more productive. We make people's knowledge compound into intelligence no one holds alone.

Science is the hardest version of the problem. The same brain serves anyone who builds knowledge with an AI beside them, from research groups to a family planning a move.

Sharing you control
Nothing leaves by default. Your brain is private until you decide otherwise.
Share a cluster, not your brain. You choose exactly what travels.
Sharing is revocable. What you granted, you can withdraw.
Credit travels with every claim. Whoever established it stays named.

Everything here works today for one brain. The network it plugs into is what we are building it for.

Sapience turns what we each know into what we all know. Owned, never taken.

05   The idea

What if it learned your work, for you alone?

The way it works today: your work trains a frontier model, then stays locked inside one vendor's tool, with nothing flowing back to you. The smarter each of us gets alone, the more fragmented the whole becomes.

You own your intelligence, captured once and portable across every AI. Because it stays yours, signed and credited, it can connect.

The bind

Give your work away

The AI that knows your field exists because a lab trained on work like yours. It learns from you, for everyone but you.

Or keep it, and get a generic tool

Every session starts from zero

Opt out and your data stays yours, but the model stays frozen: no memory of your methods, your decisions, your last six months.

The way out

Learning that is yours

Sapience learns continually from your work, for you alone. Nothing trains a foreign model. When you choose, it connects with others' knowledge, credited. When small open models are enough, the whole system runs on hardware you own.

Ownership and connection are not opposites. Each needs the other. Only what stays yours can connect without handing over the underlying work, and the more it connects, the more it is worth owning.
The only superintelligence that matters is personalized superintelligence.
06   The horizon

What if your expertise paid you, and not a broker?

An enclosure of knowledge is underway: frontier labs bake expert work into private weights and rent it back. What matters next lives in experts' heads and private work, on the order of a million times larger than the web (Trask and Strahm, Institute for Progress, 2025). Whoever makes private expert knowledge usable without exposing it defines the next era of AI.

Ownership has an endgame: because every claim is signed to its creator, expert knowledge can one day be licensed directly, from the people who make it to the labs that need it, with no broker taking the difference. The attribution registry is the rails.

Attention appeared in 2014, in an academic lab. The protein-folding breakthrough built on it landed six years later, somewhere else entirely. The intelligence existed. The connection was the bottleneck.
We are deciding whether the next era of intelligence is rented from five labs, or owned by the people who create it.

That is the gap Sapience Labs exists to close: for one mind, and between all of them.

07   Where it stands

What people say.

“This is something incredibly useful. The cross-domain connections are things I would never have found on my own.”
Cosmologist · University of Cambridge
“The first two papers it surfaced were spot on for my research. It is giving me ideas on what to do next.”
Experimental physicist · CERN
“I one billion percent see the benefit. The more collaboration across the world, the faster science progresses.”
Researcher · Queen Mary University of London

The team. Physicists, engineers, and researchers from Berkeley, Cambridge, Harvard, SpaceX, Google, DeepMind, and OpenAI, quietly building; more than $100M raised for previous ventures. If this is the problem you want to spend your years on, come build it with us.

6
Patent families
Every AI
Claude · GPT · Gemini · open
Yours
On your key · we never train on it

We keep only the knowledge objects we extract, never your raw conversations, and we never train on your data. The collective network is built from shipped primitives and a roadmap; we describe what is live as live. The headline results are further down this page; the full protocols and run files go to partners.

08   Open at the core

Open standard, sealed engine.

The Knowledge Object is designed as an open standard: signed, credited, typed, interoperable with existing open knowledge formats. Because the format is open, your knowledge is never hostage to us. What stays ours is the intelligence that connects it.

Open · the independence
  • The Knowledge Object standard: signed, attributed, typed-edge format
  • A free client: capture your work across any AI, kept locally, on your key
  • Interoperable with open knowledge formats, not a walled schema
Proprietary · the moat
  • Cross-person and cross-organization matching
  • The importance-weighting and property taxonomy that rank what matters
  • The consolidation and learning engine, the hosted network, the attribution registry

The open-core company for learning you own. The Knowledge Object standard is stewarded by CoreTx, our sister company, so the format your knowledge lives in is never hostage to us either.

sa·pi·ence n. 1. wisdom; the trait our species is named for. 2. knowledge grounded in experience, compounded when minds connect.

Not a smarter individual. A collective that gets wiser.

A brain for each of us. The beginnings of a collective one.

Read the declaration ›

09   Begin

We are opening a small number of beta places each day.

Ask for a place and we send an invite code. The build takes about ten minutes: no setup, no API key, one line builds your brain from your recent work and gives you a private link. The first thousand builders get the network free, for life. The first discoveries will be theirs. Come for the independence. Stay for the network.

Prefer email? oliver@coretx.ai

$ curl -sL "https://spnc.ai/install?i=YOUR-CODE" | bash copy
your code arrives with your invite

Once your code arrives: Mac and Python 3, your history is read locally, the asking and connecting run on us. Not a terminal person? There is a one-click Mac installer. Prefer to see one first? Ask a live example brain your own question.

read on your laptop, never uploaded
Live sync with your AI

Connect over MCP and the loop closes: your AI reads what you know mid-conversation and writes what you learn back, so every session starts smarter than the last.

search(query) // find it in your knowledge
load_context(topic) // start the session smarter
remember(insight) // signed to you
ask(question) // a cited answer
Privacy

Your link is private and unguessable. During the beta, hosted demo brains auto-delete after 14 days of inactivity; a brain on your own machine is yours for good. For a fully private or on-prem deployment, talk to us.