HumanSignal

Data Generation & Labelling

HumanSignal develops data labeling and AI evaluation infrastructure for machine learning teams.

Reviewed by 7wData
Share Email

On this page

Profile

Open-source data labeling platform and managed data services for AI teams to annotate, evaluate, and fine-tune machine learning models.

HumanSignal develops data labeling and AI evaluation infrastructure for machine learning teams. The company was founded in 2019 as Heartex by Michael Malyuk, Maxim Tkachenko, and Nikolay Lyubimov, who met while climbing in the Himalayas and identified a fundamental challenge in AI: access to high-quality labeled data. The flagship product, Label Studio, is an open-source data labeling platform that has grown to serve over 250,000 data scientists and annotators globally, who have labeled more than 200 million pieces of data.

In June 2023, Heartex rebranded to HumanSignal to reflect its expanded mission beyond labeling toward human-in-the-loop AI evaluation and fine-tuning. The company raised $25 million in Series A funding in May 2022, led by Redpoint Ventures, bringing total capital to $30 million. Recent product evolution includes LLM evaluation capabilities introduced in June 2024, multimodal agent evaluation launched in June 2025, and a strategic acquisition of Erud AI announced in November 2025.

This acquisition marked HumanSignal's expansion into managed data services, establishing physical data labs where domain experts and annotators create purpose-built datasets tailored to specific use cases. Named customers include Outreach, Zendesk, Bombora, Geberit, and Yext, with enterprise deployments also reported at Bayer, EA, Epic Games, Lufthansa Technik, and Verily. The company maintains approximately 76–87 employees as of early 2026 and operates with a dual business model: a free and open-source community edition paired with Label Studio Enterprise and newly launched HumanSignal Services for managed data creation and evaluation.

Track HumanSignal and 240+ vendors.

335k+ subscribers read the daily AI & data note. One email, both newsletters. Unsubscribe anytime.

Who buys this

  • Enterprise AI teams building custom models and fine-tuning foundation models
  • Machine learning platforms and infrastructure providers needing annotation workflows
  • Frontier AI research labs requiring high-quality, domain-specific datasets
  • Autonomous systems and robotics companies labeling multimodal sensor data
  • Defense and government contractors with on-premises data annotation requirements

Publicly disclosed clients

  • Outreach
  • Zendesk
  • Bombora
  • Bayer
  • EA (Electronic Arts)
  • Epic Games
  • Verily

Strengths and what to watch

Strengths

  • Open-source foundation provides security and auditability; enterprises can self-host and avoid vendor lock-in, a structural advantage competitors with closed-source models cannot match
  • 1M+ practitioners in the community using Label Studio, creating network effects and reducing adoption friction for enterprise tier-up
  • Erud AI acquisition adds managed data services capability, enabling end-to-end "data lab" model that combines software, infrastructure, and domain expertise

Watch for

  • No Series B funding announced since May 2022; $30M total raised is modest for 4+ years of operations, suggesting potential runway or profitability constraints worth monitoring
  • Competitive market with well-funded entrants (Scale AI, Labelbox, Snorkel) moving into enterprise; must maintain velocity on product innovation to defend positioning
  • Erud AI acquisition's success depends on scaling managed services profitably; margins and customer concentration in services arm are key metrics to track

Recent moves

Key Information

Industry
Data Generation & Labelling
Founded
2019

Frequently Asked Questions

What is HumanSignal?

HumanSignal is an open-source data labeling platform and managed data services company founded in 2019. Its flagship product, Label Studio, helps AI teams annotate and evaluate machine learning models. The platform serves over 250,000 data scientists globally who've labeled more than 200 million pieces of data.

What is Label Studio?

Label Studio is HumanSignal's open-source data annotation tool that lets teams label data for machine learning models. It's used by over 1 million practitioners in the community. The platform supports various data types and integrates with ML workflows, with enterprise versions available for larger organizations.

How does HumanSignal evaluate AI models?

HumanSignal introduced LLM evaluation capabilities in June 2024 and multimodal agent evaluation in June 2025. These tools let teams assess and iterate on AI systems using human feedback. The platform enables fine-tuning foundation models with domain-specific labeled data for improved model performance and accuracy.

What are HumanSignal Services?

HumanSignal Services is the company's managed data creation offering launched after acquiring Erud AI in November 2025. It operates physical data labs where domain experts and annotators create purpose-built datasets tailored to specific AI use cases. This extends HumanSignal beyond software into full-service data preparation.

Which companies use HumanSignal for AI labeling?

Named customers include Outreach, Zendesk, Bombora, and Bayer. Enterprise deployments also appear at EA, Epic Games, Verily, and Lufthansa Technik. These organizations use HumanSignal's platform to label data and fine-tune machine learning models for production AI systems across various industries.

Is Label Studio open source?

Yes, Label Studio is open-source, allowing enterprises to self-host and avoid vendor lock-in. This open-source foundation provides security, auditability, and structural advantages over closed-source competitors. HumanSignal offers both a free community edition and Label Studio Enterprise with additional features for larger organizations and compliance requirements.

How HumanSignal compares

Direct head-to-head against 3 competitors. Picked by 7wData.

This company

HumanSignal

Positioning
Open-source data labeling platform and managed data services for AI teams to annotate, evaluate, and fine-tune machine learning models.
Customer segments
Enterprise AI teams building custom models and fine-tuning foundation models
Strengths
Open-source foundation provides security and auditability; enterprises can self-host and avoid vendor lock-in, a structural advantage competitors with closed-source models cannot match
Watch for
No Series B funding announced since May 2022; $30M total raised is modest for 4+ years of operations, suggesting potential runway or profitability constraints worth monitoring
Recent moves
Label Studio Enterprise launches on AWS Marketplace, expanding distribution and reducing buyer friction

Labelbox

Positioning
Enterprise data labeling and AI training platform targeting frontier AI labs and large enterprises via Alignerr managed expert network.
Customer segments
Enterprise ML engineers and AI product teams at tech companies, frontier AI labs, and regulated industries needing training data.
Strengths
Alignerr network of 1M+ domain experts delivers human-in-the-loop training data, reducing workforce recruitment burden for enterprise annotation teams.
Watch for
No disclosed funding round since January 2022 raises vendor stability concerns for enterprise procurement and multi-year contract decisions.
Recent moves
Acquired Upcraft (February 2026) to automate Alignerr expert outreach using agentic sales automation.

Scale AI

Positioning
High-volume AI training data provider for frontier AI labs and US government; valued at $29B after Meta's June 2025 investment.
Customer segments
Frontier AI labs, US defense and intelligence agencies, large enterprises requiring volume human feedback for proprietary model training.
Strengths
Government and defense data annotation at volume; cleared workforce handling sensitive post-training workloads self-hosted tools cannot reach.
Watch for
Meta's June 2025 49% stake prompted Google, OpenAI, and xAI to reduce or cut data contracts, concentrating revenue risk.
Recent moves
Meta acquired 49% non-voting stake for $14.3B in June 2025, valuing Scale AI at $29B.

Snorkel AI

Positioning
Programmatic data labeling and AI evaluation platform; pivoted toward expert Data-as-a-Service following $100M Series D in May 2025.
Customer segments
Enterprise ML teams in technology, financial services, and healthcare deploying models requiring custom training and evaluation datasets.
Strengths
Programmatic weak supervision reduces manual labeling volume, lowering dataset creation costs for teams with large-scale annotation requirements.
Watch for
Multiple product pivots and leadership instability documented in employee reviews raise roadmap continuity concerns for multi-year enterprise deployments.
Recent moves
Raised $100M Series D at $1.3B valuation (May 2025) and launched Snorkel Evaluate plus Expert Data-as-a-Service products.

Sources

  1. humansignal.com — Company mission, product portfolio (Label Studio, HumanSignal Services), community scale (1M+ practitioners)
  2. www.crunchbase.com — Series A funding details ($25M, May 2022), lead investor (Redpoint Ventures), total capital raised ($30M), employee count (51-100)
  3. humansignal.com — Named customers: Outreach, Zendesk, Bombora, Geberit, Yext, Scoutbee, Welocalize; case studies confirming usage
  4. www.featuredcustomers.com — Enterprise customers and case studies including Bayer, EA, Epic Games, Verily, Lufthansa Technik as AI platform teams using Label Studio Enterprise
  5. humansignal.com — Rebranding date (June 2023), founder names, strategic rationale for expansion into evaluation and fine-tuning
  6. labelstud.io — Erud AI acquisition (November 5, 2025), HumanSignal Services launch, strategic vision for managed data labs
  7. humansignal.com — Series A funding round details, investor participants, company goals around data quality and automation
  8. labelstud.io — 2024-2025 product updates (ML backend enhancements, interface redesign, evaluation templates, multimodal agent support), community engagement metrics