How to read the map
The categories below are an editorial grouping by service model, not a ranking by valuation or quality. Workforce-country coverage is often project-specific; headquarters are not used as a proxy for where contributors work.
Alignerr is identified as a Labelbox-powered platform, so the entries should not be counted as independent parent companies. The map deliberately omits unsourced workforce counts, funding figures, and claims about which payment provider a business uses. Review the scope directly with each vendor.
The company and platform map
Each official source is linked below the table. Where an exact workforce-country list was not established from the source, the table states the question to confirm rather than inventing coverage.
| Company / platform | Service grouping | Publicly described focus | Workforce-country scope |
|---|---|---|---|
| Mercor | Expert networks | Human data, evaluations, and RL environments | Global expert network; confirm project eligibility |
| Surge AI | Human data services | Labeling workforce and tools | Confirm project-specific sourcing |
| Scale AI | Data and applications | Training data and AI systems | Confirm project-specific sourcing |
| Turing | Expert networks | Expert contributions and AI development | Confirm project-specific sourcing |
| Handshake AI | Expert networks | AI training projects for experts and generalists | Confirm program eligibility |
| Deccan AI | Human data services | Expert-curated datasets and model evaluation | India sourcing described publicly |
| micro1 | Expert networks | Training data and agent evaluation | Confirm project-specific sourcing |
| Invisible | AI operations | Applied AI and human-in-the-loop operations | Confirm project-specific sourcing |
| Rise Data Labs | Human data services | Data labeling and expert recruitment | Confirm project-specific sourcing |
| Luel | Data collection and licensing | Rights-cleared multimodal data | Global collection network described by company |
| Prolific | Research participants | Research and human-feedback participant access | Eligibility varies by study and participant location |
| Alignerr | Expert network / Labelbox | Expert contributions to AI training | Confirm current role eligibility |
| DataAnnotation | Contributor platform | Expertise-based AI training work | Confirm current role eligibility |
| Sama | Managed annotation | Data annotation and labeling | Confirm delivery location for the project |
| CloudFactory | AI operations | AI implementation and human data workflows | Confirm delivery location for the project |
| Appen | Human data services | AI training data and human evaluation | Confirm sourcing and project eligibility |
| TELUS Digital | Human data services | Data for AI training | Confirm sourcing and project eligibility |
| Toloka | Human data services | Training data for agents and language models | Confirm project-specific sourcing |
| Innodata | AI data services | Generative and traditional AI data services | Confirm project-specific sourcing |
| TaskUs | AI data services | AI data operations and managed services | Confirm delivery location for the project |
| Labelbox | Data platform and services | AI data workflows and expert services | Confirm expert-network eligibility |
| SuperAnnotate | Data platform and services | AI data infrastructure | Confirm project-specific sourcing |
| Kili Technology | Annotation platform | Annotation tooling and workflows | Confirm whether workforce services are in scope |
| Keymakr | Managed annotation | Image/video annotation and data collection | Confirm project-specific sourcing |
| Shaip | AI data services | Data collection, annotation, and evaluation | Confirm project-specific sourcing |
Build your workforce or buy a managed service?
A managed data project can bundle sourcing, tools, quality control, and delivery. Building your own network gives you more responsibility for recruitment, agreements, identity checks, work allocation, quality, and payments. These are different operating models even if both produce the same kind of dataset.
Compare the retained work, not only the quoted cost per label. Ask whether you can audit rights and consent, reproduce acceptance decisions, and reconcile contributor compensation to delivered work. If you operate the network yourself, contracting and payout infrastructure becomes part of your own system.
What to ask each provider
Use a project brief that names the modality, expertise, languages, country constraints, acceptance criteria, and intended use of the data. Then ask each provider to explain what is included and which obligations remain with your company.
- Who recruits and contracts with contributors?
- What rights and consent evidence accompany delivery?
- How are quality disputes and rework handled?
- Which countries are eligible for this specific project?
- Who holds contributor records and how can required evidence be accessed?
- What changes the price: expertise, collection difficulty, review effort, or volume?
Keep the comparison dated
This research snapshot was checked on 13 October 2026. Company offerings and eligibility can change; re-check the linked sources before procurement. A quarterly editorial refresh is recommended, but no automatic update schedule is implied.
Toku provides contributor contracting and payment infrastructure. It is not presented here as a substitute for a data-labeling company’s sourcing expertise or annotation product.
General information for teams planning contributor programs. Classification, tax, payment eligibility, and data obligations depend on the actual arrangement and applicable local requirements.