Fundraising Waitlist open · launching 22 Nov 2026

AI-powered research, built by everyone. Credited fairly.

Every field’s next breakthrough will be AI-powered. CrowdSota lets anyone push the state of the art in any field — biology, materials, climate, medicine, math, physics, social science, and AI itself — with their own tokens, compute, specialized tools, and ideas. A fair-credit ledger records who did what.

View the pitch deck · Try the credit simulator

Any fieldBiology to math to AI — wherever AI can do research
4Ways to contribute: tokens, compute, tools, ideas
2Equal front doors: Open and Enterprise
14CRediT contributor roles (ANSI/NISO standard)

Any field

State of the art, in every field — pushed by the crowd.

SOTA means state of the art in whatever field you care about, not just AI benchmarks. Pick a field to see what a CrowdSota challenge could look like — in the open, and inside a company.

Example challenges  Illustrative only — these are not live projects, partners, or results.

The problem

The work can run all year, in every field. Credit, and the hackathon, do not.

The same gap shows up in public research and inside companies: people and AI agents can already do research work, but credit, access, and funding still sit in closed rooms, or in a once-a-year burst.

In open science

Closed rooms, thin credit

AI agents can already search literature, analyze data, run experiments, and draft papers — in biology, materials, math, and beyond. The tools often live in a private workspace. Decentralized networks may reward GPUs or tokens, which is not authorship. Open-science platforms welcome people, but are not built for bring-your-own keys or agents on a public ledger.

Inside companies

Weekend toys, or one heavy week a year

Internal hackathons are either lightweight individual exercises that do not take on large problems, or serious annual events that are effort-heavy and then go dark. Teams work in silos. There is often no direct, ongoing incentive, and leadership cannot see how AI spend maps to R&D goals — across, for example, a company with 100 teams of about five people working independently.

Two front doors · equal weight

One core. Two front doors.

The same fair-credit ledger and AI research loop power both. Switch between them.

CrowdSota Open

A public lab for every field, aimed at published state of the art.

Anyone can participate. Contributions of tokens, compute, specialized tools, and ideas go on a public ledger, with fair credit that can become authorship. The long-term aim is crowd-sourced, AI-powered research that reaches published state of the art in its field — we will not claim a SOTA result until one exists.

01

Open participation

You do not need to be at a frontier lab or a big institute. Researchers, students, contributors, and sponsors in any field can join a public challenge and be named for the work they actually did.

02

A public contribution ledger

Effort (tokens, GPU-hours, runs) and impact (runs that move the field’s stated state-of-the-art metric) are both visible. When a project is written up, authorship follows CRediT roles.

03

Toward published results

The point is not a token balance. It is a shared, checkable record of research that can be reviewed, reproduced, and, if it earns it, published — in that field’s own venues.

01 · IDEAQuestionSomeone proposes a problem in their field.
02 · RUNSAgents + computeBYOK tokens and GPU workers run the experiments.
03 · CHECKSpot-checksA sample of runs is re-executed elsewhere.
04 · IMPACTState of the art movesRuns that move the field’s metric earn impact credit.
05 · WRITE-UPAuthorshipNamed CRediT roles on the published result.

How it works

Four ways to contribute. One ledger. Any field.

The loop is the same in CrowdSota Open and CrowdSota for Enterprise. People contribute tokens, compute, specialized tools, or ideas; AI agents do research work; a fair-credit ledger records it. In Open, that ledger is public. In Enterprise, it is visible across team lines, inside the company.

Tokens

Bring a key, or fill the pool

Use your own model API key for a single run (BYOK), or donate credits to a shared pool that others can draw from. The ledger records which path funded the work.

Compute

Run a worker on your GPU

A lightweight worker picks up queued experiments — simulations, model training, data analysis. Samples are spot-checked by re-running them elsewhere, so a contribution is not just a claim.

Specialized tools

Share the tools your field runs on

Give agents access to the standard tools of your industry — simulation software, wet-lab or EDA tools, protein-folding or chemistry packages, domain datasets, licensed instruments or APIs. The ledger credits who made the work possible.

Ideas, code, reviews

Shape the science

Propose a question, ship code, or review a result. Domain experts steer and judge; agents execute; both are attributed.

contributeagents runspot-checkstate of the art moves?credit + authorship / funding

Try it

See how your contribution would be credited.

Pick a field and a way to contribute, then move the sliders. The split is recalculated against an example project with five fictional contributors.

Illustrative
Field
I want to contribute
Source
0%your share
    Effort0
    Impact0
    Total0
    CRediT role(s) on the write-up

    Illustrative simulator — not a real scoring formula and not live data. Impact is weighted more than effort to show the idea; the real weights are still to be designed.

    Fair credit

    Two layers, then authorship — in public or across the company.

    Credit is not a vibe and it is not only a token balance. We separate measurable effort from evidence of impact, and we map people (and agents) onto standard contributor roles when a result is written up — in any field.

    Layer 1

    Effort

    Tokens spent, GPU-hours contributed, specialized tools and datasets provided, runs completed. Who showed up, and with what resources?

    Layer 2

    Impact

    Runs that move the stated state-of-the-art metric count more than runs that do not. Whose work changed the result?

    Write-up

    Named roles

    Authorship follows the CRediT taxonomy (ANSI/NISO): Conceptualization, Software, Validation, Writing — not one undifferentiated byline.

    Example — not live dataIllustrative only. Fictional contributors.
    ContributorTypeEffortImpactCRediT role
    A. ChenideasHypothesis + protocolAccepted into challenge #1Conceptualization
    @gpu-node-14compute18.4 GPU-hours, 12 runs3 runs improved the scoreInvestigation
    R. OkonkwotokensBYOK, 4 agent reviewsCaught a failed spot-checkValidation
    credit-pooltokensShared credits, 2 runsNo score moveResources
    L. MoreautoolsFolding package + curated assay datasetUsed in 4 validated runsResources, Software

    Names and figures in the table are fictional examples to show how the ledger would read. They are not traction.

    Why now

    The pieces exist. The continuous, open lab does not.

    The observations below are from sources we reviewed in November 2026, not market-size claims. The enterprise framing is founder experience, not a survey.

    01

    Autonomous research is entering the scientific record.

    Sakana’s AI Scientist work was published in Nature in March 2026, including a discussion of credit, watermarks, and the limits of generated papers.

    sakana.ai/ai-scientist-nature ↗
    02

    Researchers are already buying agent workbenches — mostly private ones.

    2026 Product Hunt launches such as OpenScience, Meshia, and ClawsGO Science show demand for AI research agents and, for Meshia and OpenScience, bring-your-own keys or wallets. They are not public contribution ledgers.

    OpenScience ↗Meshia ↗ClawsGO Science ↗
    03

    Permissionless compute contribution has been demonstrated.

    Prime Intellect’s INTELLECT-2 described a globally distributed RL run in which heterogeneous machines could contribute, with verification of inference work.

    primeintellect.ai/blog/intellect-2 ↗
    04

    Contributor roles already have a standard.

    CRediT is an ANSI/NISO taxonomy of 14 roles, used across disciplines, so research outputs can name kinds of work, not only an ordered author list.

    credit.niso.org ↗
    05

    Inside companies, the serious version is still annual.

    From the founder’s time at Amazon and Toyota: many internal hackathons stay individual and toy-scale; the serious ones fund real products, but only once a year and at high operational cost. A year-round loop with a fair-credit ledger is the gap we want to close. This is lived experience, not a market statistic.

    How we’re different

    One core. Two gaps we have not seen filled together.

    For CrowdSota Open: open participation in any field, BYOK plus a shared credit pool, specialized tools as credited contributions, agents that do research work, and a public fair-credit ledger. Adjacent products cover some of these, not all. This comparison is from a November 2026 review of public pages, not a paid landscape study.

    CrowdSota plannedMeshia / OpenScienceSakana / FutureHouse / Autoresearch / OrikataBittensor / Prime IntellectResearchHub / Zooniverse
    Open participationYes — public lab, any fieldPrivate team workspacesSingle-team systemsOpen for compute / miningOpen to people
    BYOK + shared creditsYes — key per run or poolBYOK / wallet / credits in-productTypically hosted or single-teamCompute and token rewardsNot built around model keys
    AI agents doing researchYes — queued experimentsYesYesTraining / inference, not paper authorshipHuman review or classification
    Specialized tools as credited contributionsYes — software, datasets, instruments, APIsNot found as a credited contributionNot found as a credited contributionNot found as a credited contributionNot found as a credited contribution
    Provable fair creditPublic ledger + CRediTNo public contribution ledger foundOrdinary internal or paper creditEmissions for work, not CRediT authorshipOpen, but not agent + BYOK credit

    For CrowdSota for Enterprise, the contrast is not that public table. It is the difference between a weekend or annual hackathon and a continuous, funded R&D challenge with a ledger that crosses team lines. We are not naming other internal-innovation vendors here; we have not done a paid landscape study of that category.

    Business model proposed

    Enterprise, the open platform, and sponsored challenges.

    These are proposed revenue lines. We have no revenue yet.

    01

    Enterprise subscriptions & pilots

    Always-on internal R&D challenge platform, starting with paid pilots tied to a real leadership goal.

    02

    Fee on pooled credits

    A platform fee on pooled credits on the open platform — not on a contributor’s own BYOK key.

    03

    Sponsored challenges

    Organizations sponsor public state-of-the-art challenges in the fields they care about.

    Roadmap

    Stand up the core. Open a public challenge. Pilot inside a company.

    22 November 2026 · now

    This site and the waitlist

    A public pitch, a pledge form, and conversations with researchers, contributors, sponsors, investors, and companies interested in a pilot. The product is not live yet.

    Next proposed

    First open challenge

    One narrow, measurable public task with a clear state-of-the-art metric. Waitlist pledges of keys, GPU-hours, specialized tools, and ideas are invited to join.

    Then proposed

    The ledger, workers, and spot-checks

    A visible log of tokens, GPU-hours, and runs — including BYOK versus pool. GPU workers pull queued experiments. A sample of runs is re-executed to catch bad results.

    In parallel proposed

    First enterprise pilot

    An always-on internal challenge tied to a leadership R&D goal and a budget, with credit visible across team lines. Top work can be put forward for product funding.

    Enterprise needs proposed

    SSO, private deployment, audit trail

    The controls companies will need before this runs on internal work.

    Aim proposed

    Impact credit, write-ups, funded products

    Open results authored with CRediT roles; the long-term aim is published state of the art across many fields, claimed only if it exists. Inside a company, the best work is funded as a real product.

    Pitch deck

    Twelve slides. Right here.

    Use the arrows, the ← → keys, swipe, or go fullscreen. Placeholders for the raise and the founding team are marked in amber. Open and Enterprise have equal weight.

    CrowdSota

    AI-powered research, built by everyone. Credited fairly.

    Every field’s next breakthrough will be AI-powered. CrowdSota lets anyone push the state of the art in any field with their own tokens, compute, specialized tools, and ideas — on a fair-credit ledger.

    Two equal front doors: CrowdSota Open and CrowdSota for Enterprise.

    Pitch · 22 November 2026

    01 / 12

    Problem

    The work can run all year. The credit system cannot.

    • In every field, AI agents can already do research work — but labs, keys, and credit often stay closed.
    • Inside companies: hackathons are weekend toys, or serious but once a year.
    • Silos keep ideas, compute, tools, and credit inside one team. Leadership cannot see how AI spend maps to R&D goals.

    CrowdSota

    02 / 12

    Solution

    One core. Two front doors, equal weight.

    • The core — a fair-credit ledger and an AI research loop for any field: tokens, compute, specialized tools, ideas.
    • CrowdSota Open — a public lab, a public ledger, authorship via CRediT, aimed at published state of the art in biology, materials, climate, medicine, math, physics, social science, and AI.
    • CrowdSota for Enterprise — always-on AI-powered R&D challenges across every domain a company works in: leadership goals and budgets, credit across team lines, top work funded as products.

    CrowdSota

    03 / 12

    How it works

    Tokens. Compute. Tools. Ideas. Same loop, any field.

    • Tokens — bring an API key per run, or donate to a pool (or a company budget).
    • Compute — run a worker; queued jobs; sample re-runs for checks.
    • Specialized tools — simulation software, wet-lab or EDA tools, folding or chemistry packages, domain datasets, licensed instruments or APIs.
    • Ideas / code / reviews — domain experts steer, agents execute, both are named.
    • In Open the ledger is public. In Enterprise it is visible across team lines.

    CrowdSota

    04 / 12

    Credit model

    Effort, then impact, then named roles.

    • Effort — tokens, GPU-hours, tools and datasets provided, completed runs.
    • Impact — runs that move the field’s stated state-of-the-art metric, or a company goal, count more.
    • Authorship / recognition — CRediT roles on a public write-up, or visible credit across teams inside a company.
    • Spot-checks make compute claims testable.

    CrowdSota

    05 / 12

    CrowdSota Open

    A public lab for every field, aimed at published state of the art.

    • Anyone can contribute tokens, compute, specialized tools, or ideas — to challenges in biology, materials, climate, medicine, math, physics, social science, or AI.
    • Every contribution is recorded on a public ledger, with fair credit.
    • When a result is written up, authorship follows CRediT roles.
    • The aim is crowd-sourced research that can be published. We will not claim SOTA until it exists.

    CrowdSota

    06 / 12

    CrowdSota for Enterprise

    From a weekend or an annual event to a year-long R&D challenge.

    • At Amazon, as the founder experienced it: internal hackathons were individual, weekend-scale “toy” exercises — not a way to solve large problems.
    • At Toyota, as the founder experienced it: the annual hackathon is serious, and the top five products get funding — but it is once a year, and effort-heavy.
    • CrowdSota automates that into a continuous, year-long hackathon for AI-powered R&D in any domain the company works in, with direct incentives, across silos.
    • Leadership sets goals and budgets, and sees how AI spend maps to those goals. Amazon and Toyota are prior workplaces, not customers.

    CrowdSota

    07 / 12

    Why now

    Every field is becoming AI-powered. The open lab is not here yet.

    • AI-scientist systems are in the literature (Sakana, Nature, March 2026).
    • BYOK research workbenches launched in 2026 (e.g. Meshia, OpenScience).
    • Permissionless compute contribution has been shown (INTELLECT-2). CRediT already standardizes roles across disciplines.
    • Inside companies, the serious format is still typically annual — from founder experience, not a market survey.

    CrowdSota

    08 / 12

    Landscape

    Adjacent in two markets. Equivalent in neither.

    • Open: Meshia / OpenScience — agents + BYOK, private workspaces. Sakana / FutureHouse / Autoresearch / Orikata — single-team scientists. Bittensor / Prime Intellect — compute rewards. ResearchHub / Zooniverse — open, not agent + BYOK ledgers.
    • Enterprise: the contrast is weekend or annual hackathons, and internal work that does not carry fair credit across team lines.
    • We have not done a paid landscape study of internal-innovation vendors, so we do not name a competitive set there.

    CrowdSota

    09 / 12

    Business model · proposed

    Enterprise, the open platform, and sponsored challenges.

    • Enterprise subscriptions and paid pilots.
    • A platform fee on pooled credits on the open platform (not on a contributor’s own BYOK key).
    • Sponsored state-of-the-art challenges, in any field.
    • These are proposed. We have no revenue yet.

    CrowdSota

    10 / 12

    Go-to-market · proposed

    One public challenge. One enterprise conversation.

    • Open a waitlist for researchers in any field, contributors, sponsors, investors, and companies.
    • Collect pledges of tokens, GPU-hours, specialized tools, and ideas for a first, narrow public challenge with a public ledger.
    • In parallel, talk to companies about a year-round internal pilot tied to a real leadership R&D goal.
    • Enterprise controls — SSO, private deployment and data controls, an audit trail — are on the roadmap.

    CrowdSota

    11 / 12

    The ask

    We are raising [Raise amount TBD]

    To fund the ledger, the first open challenge, early enterprise pilots, and a small product-led team. Details of use of funds will follow the amount.

    Team: [Founders]

    CrowdSota

    12 / 12

    01 / 12

    Waitlist & pledges

    Reserve a place — in the open lab, or for a company pilot.

    Tell us who you are, your field, and what you might contribute — or if you want to talk about an enterprise pilot. This is a pledge of interest, not a contract and not a sale of tokens or equity.

    Investors

    Talk to us about the raise

    The button opens the waitlist form with your role set to Investor.

    Request a conversation
    Companies

    Run an always-on R&D challenge

    Interested in a pilot tied to a real leadership goal, in your own domain?

    Ask about a pilot

    This is a pledge of interest, not a binding commitment.