DAACWhat you get

Built for markets too small to matter before.

Digital tools capture the field and hubs open the public record; every solution here stands on that supply. What a buyer takes away is data they can act on, a system built to run their own operation, or analysis already turned into something they can hand to a board or a regulator.

Data access · what a buyer can query or connect to
Spec · not running yetfrom

Data marketplaces

Priced, permissioned access to reconciled data: a query surface, an API, or an MCP feed (the protocol a buyer’s own AI agents connect through), or an integration into their systems. Spine is the exchange underneath; it stands at spec today.

Better access to data lets a buyer underwrite an investment they could not justify before, and measure the impact of the one they already made.

Risk managementImpact measurementDue diligenceOpportunity assessment
Who pays for it →
Live todayfrom

Sector hubs

Open, source-rated reference hubs for opaque sectors, a public good, a front door, and public-data supply.

All the data on a topic, scattered across the public internet, ends up in one interactive hub a visitor can read across.

See the hubs →
Operational capability · what a buyer can run
Build · none live on Frame yetfrom

Digital tools

Working tools built on a framework, capturing frontier data where the work happens: the ecosystem’s first-party supply. Built on the Datastake framework today, on Frame in the DAAC-native form, or on whichever framework Forge is tasked against.

A digital tool records what is happening at the source, so a decision about the field rests on the state of the field now, not on a report written weeks later.

See the proof →
Running todayfrom

Digital system frameworks

For when the need is not one tool but a whole software system: the framework other tools are then built on, which Forge builds and rebuilds while it runs. Datastake is the worked example; Frame is DAAC’s own target framework.

Holding the framework with DAAC means it never quietly falls behind what runs on it, so the system keeps being customised as the buyer’s own needs change.

How the framework evolves →
Scoped per engagementfromForge

Customised agentic workflows

Customised agentic workflows scoped to one operational job, distinct from the multi-stakeholder digital tools above. Forge builds them the way it builds those, composed against a framework under supervision.

One operational job that used to need a standing team instead runs as a supervised agentic workflow, scoped to exactly that job.

The Tools that build them →
DeliveredfromThe operating model

Delivered HITL and headless workflows

Workflow engagements already in the field: human-in-the-loop, where a person holds the checkpoints, and headless, where the loop runs unattended once proven. Each is one instance of the same reproducible operating model.

A workflow proven under a person’s checkpoints keeps running once the loop is proven, unattended, without dropping the checkpoint it was built on.

The operating model behind it →
Communication artifacts · what a buyer can read or show
On requestfromMason

Research reports

Findings written up and sourced, held to the same corroboration standard Crawl holds its hubs to, delivered as a document a buyer can act on.

A buyer acts on findings that are sourced and corroboration-rated, not a summary they have to take on faith.

Ask about a report →
This site is onefromMason

Websites

A public front door built from config-driven templates, and re-shipped whenever the record behind it changes: the same practice stood up the door for every digital tool in the ecosystem below, and for partners beyond Datastake’s own portfolio.

A front door built from a template gets rebuilt only where the property differs, so the site follows the record instead of being rebuilt from scratch each time.

See a site built for a partner →
On requestfromMason

Presentations

The same findings, reshaped for a room: a deck built to brief a client or a partner on what the data shows.

A partner walks into the room already briefed on what the data shows, not still being introduced to it.

Ask about a deck →
From signal to solution

Every build starts from a sector’s public record, or a system somebody already depends on.

It starts from a sector’s public record, scattered but real, or from a system somebody already depends on, with its data and its edge cases intact, never from a proposition. That is what makes it buildable at all.

Hub update · Hub inception

A hub is what the scatter becomes.

Most of what is known about an opaque sector is already public and unusable: registers, portals, reports and datasets, in incompatible shapes, at unknown freshness, behind no single front door. Crawl runs a loop over that scatter and Mason stands the door in front of it, open for anyone to read.

FIG · A sector’s scatter goes in, a hub that dates its own claims comes out
  1. 01Scan
    Registers first

    Fan out across a sector’s public sources, source-anchored rather than keyword-guessed. The authoritative registers are enumerated first, then the periphery.

  2. 02Verify
    Rated by corroboration

    Every finding is rated strong, moderate, single or contested, and its sources recorded. A single-source claim is labelled as such, never dressed as corroborated.

  3. 03Organise
    The record core

    Findings are arranged into a navigable structure shaped against the suite’s primitives. Crawl emits identity, sources, corroboration, freshness and provenance, and nothing downstream overwrites them.

  4. 04Front door
    Mason’s side of it

    Mason builds the door over the record Crawl owns, and re-ships it whenever Crawl refreshes what is behind it.

  5. 05Maintain
    Dated and re-checked

    Every record carries a last-verified date and a link-health signal, and the scan is built to re-run on a cadence. Crawl’s own words: freshness is the moat.

each run folds back into the sector’s memory, tuning discovery, never the bar

App migration · App build

A digital tool arrives with its history intact.

It begins with a system somebody already depends on, and the work is to move it onto Frame without dropping anything it had learned. A production app has already made that move, checked to behave identically to the one it replaced.

FIG · The input is a running system, and what must not change is how it behaves
  1. 01Start from what runs
    Already in service

    The upstream input is an application people already depend on: its data, its users, the edge cases it learned the hard way. Convert mode takes a running system, not a brief.

  2. 02Recompose on Frame
    Forge, convert mode

    Forge maps every construct in the source system to the closest unit Frame already owns before proposing a new one, and records each mapping so the reasoning can be audited later.

  3. 03Prove parity
    Behaviour, not opinion

    Gauge proves the rebuild behaves identically to the system it replaces. The bar is need-fit, prior behaviour and data fidelity together.

  4. 04Ship
    On the build line

    The rebuilt app ships on the build line. That is where the relationship starts, not where the job ends.

a person signs off on the client brief, on anything promoted into the substrate, and on the first deploy of every app

The proofin the field, not on a roadmap

The same Tools build systems that look nothing alike.

Each one opens a channel of data that existed in no system before. Once that data is comparable and exchangeable, the people who hold it can realise value from it on their own terms.

A · Built and running in the field
B · The public hubs

A sector’s record holds up because every claim carries the evidence behind it.

Alongside the data the ecosystem captures, DAAC publishes open reference hubs: a public good, a front door to the work, and the ground a sector’s community forms on once it has one shared record to gather around. Crawl crawls the public web and re-checks them as the record moves.

● LiveCritical minerals · DRC

DRC Cobalt Hub

The independent reference for artisanal and industrial cobalt in the DRC: stakeholders, initiatives, a source-rated news feed and a library, re-checked as the record moves. A sector everyone depends on and no one had mapped.

cobalt.daac.ai  ↗
● LiveArtisanal & small-scale mining · global

ASM Hub

The reference for the systems that formalise artisanal mining worldwide: source-rated traceability, certification, formalisation and monitoring platforms, mapped across the sectors that rely on them. In FR and EN.

asm.daac.ai  ↗
● LiveEITI · ITIE-RDC

Extractives Transparency Hub

The open record of who owns what and who gets paid in DRC extractives: contracts, revenues and beneficial ownership, the 2026 Validation, and a source-rated updates feed in EN, FR and SW.

nashiriki.com/eiti  ↗
● LiveClimate & nature · Global South

Climate & Nature Hub

The link-checked front door to Global-South climate work: nature-based projects, funders, standards, developers and datasets, each carrying a freshness stamp so the record can be trusted, not just found.

climate.daac.ai  ↗
● LiveResponsible gold · mine to market

Gold Hub

An open, source-rated reference on responsible gold: the standards, rules, programmes, data and technology that govern how gold moves from mine to market, with every claim rated by the evidence behind it.

gold.daac.ai  ↗
Built, and going live next
Labour & Human-Rights Hubsupply-chain labour and human rights
Mtoto child-labour indexnames every source, and blends none
CETM Integrity Hubcritical energy transition minerals supply chains
On the map next, as the data deepens
Responsible sourcingdue-diligence standards and regulation, across commodities
Global-South agriculturesoft commodities, from smallholder to export