forgedata marketplace
A considered approach to AI data

Good models start
with the right data.

Find the context your models are missing.
Explore data, compare requirements, and build
a sourcing brief with Forge.

Explore the catalogue
Code & engineeringPhysical AISpeech & multimodal
Sculptural cobalt Forge F
More than source code.History. Context. Intent.
Built around your use case
The raw material of what comes next.
Purpose-built sourcing Evidence before access Clear licensing scopeHow it works
The catalogue

What are you building?

Start with a data category.
Availability is confirmed against your brief.

8 sourcing collectionsCurated by Forge
Code repositories
commit → contextissue → resolutiontest → evidence
Git · JSON · CSV
Source on request

Private repositories with merged pull requests, issues, tests and development context. Prioritise 100+ merged PRs or MRs.

Coding agents & evaluation
Egocentric video
MP4 · JSON · Parquet
Source on request

First-person activity and manipulation footage across residential and commercial environments.

Embodied AI & robot learning
Legacy COBOLCOB · CPY · JCL
Source on requestExclusive only

Existing COBOL applications with business logic, copybooks and documentation. Exclusive licensing only.

Legacy modernisation
Speech & audioWAV · FLAC · JSONL
Source on request

Multilingual speech, conversations and transcripts with language, accent and consent coverage.

Speech recognition & voice AI
Images & videoJPEG · PNG · MP4
Source on request

Licensed visual collections for perception, recognition and multimodal model development.

Computer vision
Text & documentsJSONL · PDF · TXT
Source on request

Domain-specific documents, structured text and question-answer collections with traceable provenance.

RAG & domain adaptation
Structured & geospatialCSV · Parquet · GeoJSON
Source on request

Time series, industrial sensor records and geospatial data sourced against a precise buyer brief.

Forecasting & spatial AI
Synthetic & evaluationJSONL · Parquet
Source on request

Synthetic scenarios and evaluation sets with generation methods and contamination checks documented.

Evaluation & robustness
For the data that’s hard to find

Your next dataset might
not be on a shelf.

Specify the use case, volume and rights you need.
Build a precise brief for a supplier conversation.

Forge sourcing brief

Built around your requirements.

01Modality & domain

02Volume & acceptance criteria

03Permitted use & licensing

04Timeline & budget

Save privately. Review before sharing.
forge / workspace