ai.market

AIM Data

AIM Data: the local-first companion that gets your data listed.

AIM Data runs on your own infrastructure. It profiles your data, scans for PII, scores quality, and writes the listing for you. You review, you publish, and your raw data stays where it is.

Install AIM DataView trust signals

Trust signals

Buyers see what actually ran.

AIM Data listings use nullable privacy and quality scores. A missing score means the scan did not run, and the marketplace says that plainly.

PII scan: passed

privacy_score is present and the trust service evaluates clean.

The seller ran AIM Data’s local PII scan and no detectable personal data was found.

PII scan: flagged

privacy_score is present and the trust service flagged possible personal data.

The seller ran the scan, AIM Data found content that may be personal data, and the seller acknowledged the flag at publish time.

PII scan: not run

privacy_score is null.

The seller has not enabled the local PII scan on this listing. We do not fake a clean signal when it is missing.

Quality scan: graded

quality_score is present.

The seller ran AIM Data’s local quality grading. The numeric score surfaces alongside the badge.

Quality scan: not run

quality_score is null.

The seller has not enabled local quality grading on this listing.

“Not run” is an honest buyer signal. The seller has not yet enabled the local PII scan on this listing. AIM Data offers a one-click scan that runs on the seller’s own machine. We do not run it for them, and we do not fake a clean signal when it is missing.

How it works

From local scan to live listing.

  1. 1

    Install AIM Data.

    Python or Docker. Local-first. Nothing leaves your machine during install.

  2. 2

    Prepare with allAI.

    AIM Data profiles structure, detects PII, scores quality, and prepares listing metadata on your infrastructure.

  3. 3

    Review and publish.

    You review the generated metadata and listing copy. Only the listing goes live; raw data stays with you.

  4. 4

    Deliver peer-to-peer.

    When a buyer purchases, ai.market issues a signed delivery token that your AIM Data instance honors.

You can also sell access to your data without selling the data.

If you operate data that other organizations want to train on but cannot have a copy of, AIM Federate is the path. You keep the data on your infrastructure. Models get trained on it inside a sealed enclave. Only aggregated updates leave the enclave; raw data never moves.

Learn about AIM Federate →

Install

Install AIM Data.

Download the AIM Data desktop app when the installer link is available, run it on the machine that can reach your data, then configure the connection locally. Current command-line install paths are below and may require Docker Desktop or OrbStack.

macOS and Linux

Run the installer from a terminal, then follow the local setup prompts.

curl -fsSL https://get.ai.market/aim-data | bash

Windows

Run from PowerShell, then configure your local data connection.

irm https://get.ai.market/aim-data/windows | iex

Full documentation is being consolidated with the AIM Data release process. Until then, keep the setup local: install the app, connect the data source, run scans, review the generated listing, then publish.

Not a seller? You might want Find Data, Run Federated Learning, or our Partner Program.