Skip to the results

Infrastructure for humans and AI to discover together.

Your Sapience grows from your work and stays with you across AIs. Share what you choose, with your name attached. We are building a shared intelligence for humanity, so people and AI can discover what none of us could alone.

Stored on our servers, encrypted on disk. What stays yours

Request a beta invite See it in the app

You can use Sapience on your own. Free during beta, with a usage budget per person.

Ask the example

Ask it yourself. It answers from the notes on this map and lights the ones it used.

Live · an example, not yours

Magnetometer drift after warm-up

Calibration offset vs temperature

Drift compensation model

Answer grounded in your notes

The drift is thermal. It follows the warm-up curve and settles once the mount reaches temperature. The compensation model from the bench run held in the field trial, and Tom's variant covers the vibration case.

from 3 notes · 2 sessions · 1 shared by Tom

live map · an example Sapience

I You work
II Your AI reasons
III Your Sapience keeps the record
IV The world tests it
A colleague’s loop

Fig. II · the discovery loop. Two people’s loops meet at the record.

Results

Results, with figures and scope

BABILong multi-hop accuracy against tokens of stored history, 4K to one billion, log axis: Sapience holds between 79.7 and 87 at every length, 82 at one billion; the models on their own, reading the whole history, fall away as it grows; a model with a 1M window at chance, 14.3, on the billion-token items. — Sapience holds between 79.7 and 87 at every length from 4K, 82 at one billion; the models on their own, reading the whole history, fall away as it grows. Figure and scoring notes

OpenAI MRCR v2, 8-needle items of 512K to 1M tokens: Sapience 98.4 against vendor-published values, with reader input cost per answer beside it — Sapience ties GPT-6 Astra, 98.4 against 96.3, reading about fifty times fewer tokens per answer. Figure and scoring notes

Current-value questions answered correctly, on the same 60 questions: Sapience 88, Sapience with an 8B reader 83, Mem0 48, the same model with git grep 27, a notes file 20, flat RAG 17. The 8B bar reads with Llama-3.1-8B; all other bars read with deepseek-v3.2 — Asked what a setting is now in a codebase that keeps changing, Sapience answers 88 percent right on the shipped engine, against 48 for Mem0 and 27 for the same model with git grep, on the same 60 questions. Figure and scoring notes

Works today Remembers at scale

Accuracy holds from 4K tokens to a billion tokens of stored history, with the model reading a few hundred tokens per answer.
BABILong qa3. The official cells run to 10M; 30M to 1B use our corpora with repeating PG-19 book filler.
Figure and scoring notes
Works today Stays current

Asked what a setting is now in a codebase that keeps changing, Sapience answers 88 percent right on the shipped engine, against 48 for Mem0 and 27 for the same model with git grep, on the same 60 questions.
Real git histories of Django, LiteLLM, huggingface_hub and vllm, answers graded against git, no AI judge.
Figure and scoring notes
Works today Works across models

The same model scores 86 with Sapience against 20 without it. On OpenAI's own long-context test, Sapience ties GPT-6 Astra while reading about fifty times fewer tokens.
BABILong at 128K, the same model: 86 through Sapience against 20 by the first-sentence rule; judge-scored on both sides, 86 against 65. Nothing is fine-tuned and no weights change.
Figure and scoring notes
Lab study Learns without forgetting

Sapience forgets 4 to 18 percent of a first task instead of 50 to 94, from 7B to a 397B model.
Synthetic two-task study on open models from 7B to a 397B mixture of experts; three seeds per rung. This is a laboratory result, not a feature shipped in the beta, and not a guarantee of perfect recall.
Figure and scoring notes

All results, with figures · Essays

Sapience Labs

Why it works

Sapience uses an approach based on how people learn. It keeps relevant information, with newer versions taking precedence over earlier ones. It continues to improve while you are away.

It is built with two speeds, the way a brain is. One part captures what you establish while you work, with its source. The slower part consolidates while you are away: what matters is kept, and what changed is marked as changed. When you ask, your AI reads a few hundred tokens of what is current instead of your whole history, so answers stay right and cheap as the record grows. The model underneath is a part you can swap; nothing is fine-tuned.

The theory is complementary learning systems. Seven patent families are filed across it, and the results are under Research.

i · two ways of learning

ii · awake and at rest

iii · recall

Sapience Labs

Today, and with Sapience

A pre-trained model is trained on information up to a cutoff and then stops learning. The memory that assistants offer today is a short list of preferences and details kept inside one product, with no sources and no sharing, so you still explain your project and its constraints again, including what you decided last Tuesday.

Today: You re-explain your work every session.
With Sapience: It keeps learning with you, in every AI you use.
Today: What you work out stays in one chat.
With Sapience: It stays with you and never trains anyone else's model. You connect it with people you choose, with your name on it.
Today: You find collaborators by luck.
With Sapience: It is starting to introduce you to people whose work fits yours. Being built

Sapience Labs

An applied research lab building networked continual learning for science.

Networked continual learning, starting with frontier science.

Knowledge back with the people who produce it, and moving between them again.

Collaborative superintelligence, as we use the term, is people and their AIs learning from each other, with each contribution credited to the person who made it.

The research

Every AI today learns once and serves one person at a time. Sapience learns continuously from your work and with the people you choose.

Sapience provides continual learning at the system level. It never stops learning from your work, with no model retraining, and nothing leaves your control.

Its unit is the Knowledge Object, one claim reasoned through once, carrying its type, its source, and its edges to the claims around it.

Sapience uses a complementary-learning architecture, with seven patent families filed across it. We are still validating performance against the strongest pre-trained baselines, and we report those numbers separately.

The research program · Research: results and charts

Works today

Sapience is available today.

One command installs it in about ten minutes, with no API key. It runs behind your Claude Code, claude.ai, Cursor, Cowork, and Codex sessions and starts learning from them. A handful of researchers already run it in their daily work.

Claude Code, claude.ai, Cursor, Cowork, and Codex share one Sapience across every session and interface. Each session knows what previous sessions learned, whichever tool you use.

See it in the app · Request a beta invite

A map of agent memory architectures

Every architecture announced this year falls into the first two categories: in context or in weights.

In context

retrieved text
records selected for similarity
does not preserve record structure or track which facts are current and which they superseded

improves one model’s performance

System-level learning

continual learning at the system level
your knowledge, signed and credited
tracks what replaced what, and when
provenance on every fact
portable across every model
owned by its creators, connecting, by choice, across minds

lets new knowledge build on what you have already established

In weights

retraining the model
a separate model per user
does not retain anything else, including who supplied a fact or when, and does not preserve a way to move knowledge out of the model

improves one model’s performance

Sapience Labs

The app

Works today

Your AI knows your work.

Ask about a decision in your work and see the notes behind the answer.

What follows is the live app, not a mockup.

Detail from the recording: the answer to which sintering temperature they settled on, with the notes it cites — Maria asks which sintering temperature they settled on. The answer shows the notes it came from.

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.
You can switch between Claude, ChatGPT and Codex while working and continue where you left off.
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.

Works today

Detail from the recording: Tom Reyes shared 2 notes, with who and when, in Maria’s Sapience — Tom shares two notes from the microscope. They arrive in Maria’s Sapience with his name on them.

Sapience Labs

The research program

All results, with figures, are under Research.

Sharing today and planned introductions

Two people, thinking together through their AIs.

Works today

A chosen connection starts when you invite a colleague into a shared project. Sapience keeps your AIs in sync.

Because knowledge stays owned and credited, Sapience matches it across people and organizations no single vendor could connect.

Being built

We make a found connection when your record and a stranger’s contain two complementary contributions to one problem and both of you would benefit from an introduction. Both people accept before anything is shared.

When you choose, Sapience connects your work with other people’s work. What you establish will reach a stranger working on a complementary part of the discovery you are working toward, credited to you, in time to be useful.

In an internal benchmark on discoveries published after the reasoning model’s training cutoff, Sapience ranked the approach that later won first in 11 of 13 problems, on a small internal set whose protocol we share with partners.

Glaciologists needed to sense the whole submerged front of a Greenland glacier. Engineering had the method, distributed fibre-optic sensing, eight years earlier, with no citation path between the fields; by similarity it ranked 22,746th. Sapience re-found the pairing among the first hundred it put forward. The glaciologists made the connection by hand in the end, so this is a connection recovered, not a discovery claimed.

Sapience works with existing laboratories.

Watch three strangers solve one problem ›

An enclosure of knowledge is underway: the big 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).

Under current arrangements, each scientist is several times more productive while their fields are measurably less connected (Nature, 2026). Much of AI-for-science aims to automate discovery and remove people from the process. Sapience’s discovery process depends on people. Each person decides what to share and recognizes a complementary contribution. What each person learns makes the next person faster.

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.

Where this is going

Each notebook stays its owner’s, on whichever AI they use. A line will cross only when its owner chooses, with both names on what it finds. Here, it will connect a glaciologist on Claude with a telecoms engineer on ChatGPT, credited to both.

Works today

What it does

Ask Sapience a question

Works today

Everyday use

A short recap of the day, with what is still open and what comes next.
Type a to-do the way you would say it, and “by Monday” turns into a date.

Recorded in the app at phone size. The lab is invented, and the waits for the model are shortened.

A short recap of the day, with what is still open and what comes next.

Type a to-do the way you would say it, and “by Monday” turns into a date.

One AI can hold the whole problem.

Agent swarms split a problem because one context cannot hold it, and the handoffs between agents lose constraints and drift goals. Sapience keeps what you have established outside the window and brings back only what an answer needs, so the problem never has to be split. Agents become optional workers you add when you want them.

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.

Nothing you established is forgotten.

Works today

The product

i A living map: Your work organizes itself into a navigable map, from a single thought to the whole of what you know.
ii Beyond the context window: Years of your work, ten million tokens of stored history, kept outside the window and brought back a few hundred tokens at a time, with accuracy and cost flat as it grows. 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 Sapience, every client: Tell one AI something and every other one knows it: Claude, ChatGPT, Cursor, Codex, Claude Code, or an open model.
vi Quietly, in the background: It works behind your Claude, GPT, Cursor, 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 (a read scales with the question rather than your whole 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, Cursor, GPT, and Codex.
viii Share and import, with credit: Share a project or your whole Sapience. On import it lights up where a colleague's knowledge meets yours, every claim credited to whoever established it.

Lab study

Proof Teach a model a second task and it usually loses most of the first; with Sapience it forgets 4 to 18 percent instead of 50 to 94, from 7B to a 397B model, and learns the new task as well or better.

Forgetting the first task after learning a second across seven models: sequential fine-tuning 50 to 94 percent, with Sapience 4 to 18 percent. — After learning a second task, sequential fine-tuning forgets 50 to 94 percent of the first. With Sapience, 4 to 18 percent, and the second task is learned as well or better. A lab result, seven models from 7B to a 397B mixture of experts. A model trained from its own record also answers tasks it never saw better, +6.01 points.

Scoring notes and scope

Keeping the first task while learning a second: sequential fine-tuning forgets 50 to 94 percent of the first task; with Sapience, 4 to 18 percent, across seven model rungs from 7B to 397B, the 397B a mixture of experts. A lab result; the second task is learned as well or better. Keeping the first task while learning a second: with sequential fine-tuning the first task layer is worn away, 50 to 94 percent forgotten; with Sapience it stands, 4 to 18 percent forgotten, and the second task is learned as well or better. (the 397B is a mixture of experts, 17B active; a lab result, three seeds a rung.)

Three kinds of work it is built for

Research

Keep a project’s findings with their sources, and the reasoning behind each decision, across every AI you use.

Knowledge work

A product lead, consultant or writer keeps a project’s decisions straight across Claude, ChatGPT and the rest, and the next session starts where the last one stopped.

Working together

Two people, each on their own AI, plan a trip or a side project. One AI can use the notes the other person chose to share, with their name on them.

Everyday use, outside the lab

Works today · illustration

Planning a trip, each on their own AI

Ana plans in Claude and Theo in ChatGPT. Theo shares his notes on the train times and the hotel with Ana. When Ana asks her AI which day they should travel, it answers from Theo’s notes, with his name on them. The people and the trip are invented.

Works today · illustration

A product lead, the same record in two AIs

Illustration: Priya’s record, in two AIs. Priya is a product lead (invented)

In Claude

“What did we decide about the pricing tiers, and why?”

Sources used: Pricing review, March 12; Customer interview notes; Tier limits draft.

In ChatGPT, the next day

The same question is answered using the same notes and sources.

We keep only the knowledge objects we extract, never your raw conversations, and we never train on your data.

What stays yours

Sharing: Share a project, or a slice of it, with someone you choose. Their AI can read that slice with your name attached until you withdraw it. Nothing is shared unless you choose.
What is stored: Sapience stores the notes it extracts, not the raw conversations they came from. With the local install, extraction runs on your computer and finished notes are sent to the service.
Hosted access: Your notes are stored encrypted on our servers and are never used to train anyone else's model.
Deleting: You can delete everything, at any time.
Export: Export is coming.

Open standard, sealed engine.

The Knowledge Object is an open standard: signed, credited, typed and readable alongside the open formats that already exist, so what you know is never hostage to us. We keep one thing for ourselves, the intelligence that finds the connections.

Open

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

Ours

Cross-person and cross-organization matching
A sense of what matters, learned from how people work
The learning engine and the hosted network, with credit that travels with every claim

CoreTx stewards the standard in the open.

What people say.

“It’s amazing to have one brain’s memory infuse all my sessions: my Claude Code and web sessions are fully aware of each other now, and they talk to my Codex sessions too, across providers. That brain is constantly compounding, and it augments my own: it proactively suggests new ideas and connections.”

Alex · Princeton University

“This is something incredibly useful. The cross-domain connections are things I would never have found on my own.”

Cosmologist · University of Cambridge

“This is very promising. Having the exact paper referenced is crucial. You know it’s the right thing, not something outdated.”

Researcher · CERN

“Having a 2nd brain that is systematic and can help me keep hold of all the loose strings associated with my work is definitely invaluable.”

Catherine · astrophysicist turned data scientist, beta user

“The more collaboration across the world, the faster science progresses. I one billion percent see the benefit.”

A researcher, Queen Mary University of London

The team: physicists, engineers, and researchers from Berkeley, Cambridge, Harvard, SpaceX, Google, DeepMind, and UChicago; more than $100M raised for previous ventures. If you want to work on this problem, join the team.

Questions or feedback: feedback@spnc.ai

Open questions

Being built
We are building a network that finds complementary work for your problem. Glaciologists waited eight years for a method engineering already had.
Lab study
In a synthetic two-task study, Sapience reduced forgetting of the first task while learning the second. This is a laboratory result, not a feature shipped in the beta, and not a guarantee of perfect recall.
Hosted access
Your notes are stored encrypted on our servers and are never used to train anyone else's model.
Lab study
We are still validating performance against the strongest pre-trained baselines, and we report those numbers separately.

Research

Essays in sequence (The argument, in order)

For backers and partners

We start with individual researchers. The intended path runs from individuals to organizations, and then to discovery partnerships. The results are under Research. For backing or partnerships, write to Oliver Zahn at oliver@coretx.ai.

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

Who it is for: Anyone with a live project who works with Claude, ChatGPT or Codex. You can start alone.
Cost: Free during the beta, with a usage budget per person.
How to join: The founder reviews every request. Approved people get an invite code.
First step: Open your connector link and connect your AI.

One minute with an invite code. It reaches you in claude.ai, ChatGPT, Cursor, Claude Code or Codex. No code yet? Ask for a place.

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 Sapience from your recent work and gives you a private link. The first discoveries will be theirs. Come for the independence. Stay for the network.

$ curl -sL "https://spnc.ai/install?i=YOUR-CODE" | bash

your code arrives with your invite · read on your laptop, never uploaded

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 the live example your own question.

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.

The four tools your AI gains

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, and nothing auto-deletes: your Sapience stays until you delete it, hosted or on your own machine. For a fully private or on-prem deployment, talk to us.

What are you working on? One line is enough.

Prefer email? feedback@spnc.ai

How your work stays yours

People learn individually and can do more by connecting what they know.

Sapience turns what we each know into what none of us could know alone. Read the declaration ›

Infra­structure for humans and AI to discover together.

Results, with figures and scope

Why it works

Today, and with Sapience

An applied research lab building networked continual learning for science.

Every AI today learns once and serves one person at a time. Sapience learns continuously from your work and with the people you choose.

Sapience is available today.

A map of agent memory architectures

The app

Your AI knows your work.

Your first week

The research program

Two people, thinking together through their AIs.

Where this is going

What it does

Everyday use

One AI can hold the whole problem.

Nothing you established is forgotten.

The product

Three kinds of work it is built for

Everyday use, outside the lab

Planning a trip, each on their own AI

A product lead, the same record in two AIs

What stays yours

Open standard, sealed engine.

What people say.

Open questions

Research

For backers and partners

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

Infrastructure for humans and AI to discover together.