karya

What karya does

Karya (legally DAIA Tech Pvt. Ltd.) is a non-profit data and AI infrastructure company founded in 2021 and headquartered in Bengaluru, India, that builds end-to-end pipelines for AI data collection, evaluation, and deployment while providing dignified digital work to low-income communities. The company's core technology is its AI-enabled Platform by Karya, which supports audio, NLP, computer vision, RLHF, and LLM evaluation tasks across 70+ languages (with depth across 22 official Indian languages), delivered through a multilingual Android app and chat-based interface. Complementary products include Samiksha (a multilingual LLM evaluation pipeline grounded in 11 Indian languages and 4 high-impact domains), Samudaye (a government of India-backed worker engagement platform), and Karya Learn (a tiered skilling initiative with a Digital Skills Passport paying workers $5-$25/hour). The platform supports specialized tools such as the Gender-Intentional Data Tool (Project Sanmati) and frameworks like Kahani for culturally grounded AI storytelling.

Karya monetizes through four streams: (1) B2B data services and custom dataset creation for frontier AI labs including Microsoft Research, Google, Anthropic, and OpenAI; (2) government contracts with bodies such as the Maharashtra State Government and Karnataka Health Promotion Trust; (3) foundation grants including a USD 2 million Bill and Melinda Gates Foundation grant (2023) and a USD 1 million Google.org grant (November 2024); and (4) platform licensing via a PaaS model piloted globally with the Mastercard Centre for Inclusive Growth. Go-to-market combines enterprise field sales to AI companies, direct-to-consumer worker onboarding through mobile app, and channel partnerships with NGOs such as Digital Green (Kenya) and Leyu (Ethiopia). Pricing for the platform's consumer-facing tiers ranges from ₹4.99/month (Basic) to ₹17.99/month (Premium), while worker-side pay ranges from $5/hour for basic voice recording to $25/hour for advanced technical language work, reportedly up to 20 times local minimum wage. As of January 2026, the platform had onboarded 134,565 workers across 5 countries, completed 65 million tasks, and distributed USD 3.2 million in cumulative wages.

Name

karya

Legal name

DAIA Tech Pvt. Ltd.

Website

https://karya.in

Company type

Private

Founded year

2021

Operating status

Operating

Headcount range

101–250 employees

Short description

Karya is a non-profit AI data and evaluation platform that pays low-income workers across India, Kenya, Ethiopia, South Africa, and Bangladesh to create high-quality Indic-language datasets, speech corpora, and LLM evaluations for frontier AI labs, governments, and foundations.

Ownership category

akta.pro rank

Product category

AI Data Annotation Platform

NAICS

Custom Computer Programming Services (541511), Translation and Interpretation Services (541930), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518)

SIC

Services-Computer Programming, Data Processing, Etc. (7370), Services-Prepackaged Software (7372), Services-Computer Processing & Data Preparation (7374)

akta.pro primary industry

NLP Developer Tools & Evaluation (prompting, testing, benchmarks) (HDAAADAK)

akta.pro secondary industries

Machine Translation & Multilingual NLP (HDAAADAD), Speech Translation & Multilingual Speech Tech (HDAAAFAE), AI Governance, Risk & Compliance (GRC) Platforms (HDAAAMAA)

Keywords

• AI data annotation
• Multilingual speech data
• LLM evaluation benchmarks
• Indic language datasets
• Ethical data labelling

Headquarters

HQ city

Kalba Devi

HQ country

India

HQ region

Asia

Offices2 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Personnel, Operations, Technology or R&D, Marketing or Sales, Others

Revenue model

1. B2B Data Services: Custom data solutions for AI companies including domain-specific transcription, localized translation, and multimodal dataset creation at scale. Clients include Microsoft Research, Google, Anthropic, and OpenAI.
2. Government Contracts: Projects with Indian government bodies including Maharashtra State Government, Karnataka Health Promotion Trust, and various ministries for AI data and evaluation services.
3. Foundation Grants: Funding from philanthropic organizations including Gates Foundation (USD 2 million grant), Mastercard Centre for Inclusive Growth, and other donors for specific projects and capacity building.
4. Platform Licensing: Platform-as-a-Service offering allowing external organizations to deploy ethical digital work projects using Karya's technology infrastructure.

Pricing tiers

Go-to-market motion3 records

Distribution channels4 records

Marketing channels5 records

Core offering

Karya operates a platform-powered AI data collection and evaluation service, engaging a workforce of 134,565 workers across 5 countries to produce speech, text, image, and video datasets in 70+ languages, with focus on low-resource Indic languages. Through its Platform by Karya and proprietary Samiksha evaluation framework, the company delivers custom data solutions, domain-specific transcription, localized translation, RLHF, and LLM evaluation pipelines to frontier AI labs, research institutions, and government bodies.

Product overview

Karya operates as a platform-plus-modules architecture providing end-to-end pipelines across data collection, evaluation, and deployment for AI systems. The core offering is the Platform by Karya, an AI-enabled system for creating and deploying data tasks across Audio, NLP, Computer Vision, RLHF, and LLM Evaluation. Complementary products include Samiksha (multilingual LLM evaluation pipeline), Samudaye (worker engagement platform), Karya Learn with Digital Skills Passport (worker skilling), and the Karya Institute (research arm). The ecosystem also includes specialized projects like the TB Q&A Bot, Kahani storytelling framework, and Indigenous Recipes Project, all leveraging the underlying data infrastructure. The company also offers data services including Generation, Annotation, and Digitization across Audio, Text, Image, and Video data types.

Differentiator

Problem solved

Functional benefit

Brands

• Karya Institute: An internal think tank to study how AI will transform the future of work across the Global South, conducting research across markets, skilling, and labour.
• Samudaye
• Samiksha
• Sanmati

Products and services

• Platform by Karya AI-enabled platform to create, customize, and deploy data tasks across audio, NLP, computer vision, RLHF, and LLM evaluation; serves enterprise AI labs, research institutions, and government bodies via direct sales and Platform-as-a-Service licensing.
• Samiksha Evaluation Framework Multilingual evaluation pipeline that tests large language models against real, everyday Indian contexts across 11 Indian languages and 4 high-impact domains (health, education, finance, legal); combines expert consultation, paid end-user contributions, and large-scale automated evaluation. Designed for AI labs needing India-contextual benchmarks.
• Samudaye Platform that brings low-income and rural digital workers into India's digital economy, enabling public and non-government entities to issue tasks with wage transparency, living-wage benchmarks, skill certification, and clear standards of engagement. Built with The/Nudge Institute for Government of India deployment.
• B2B Data Collection & Annotation Services Custom data solutions for frontier AI companies and government bodies including domain-specific transcription, localized translation, and multimodal dataset creation across speech, text, image, and video in 70+ languages. Clients include Microsoft Research, Google, Anthropic, and OpenAI.
• Platform-as-a-Service (PaaS) Licensing of Platform by Karya technology to external organizations (e.g., Digital Green in Kenya, Leyu in Ethiopia) so they can deploy ethical digital work projects in their own contexts with Karya's infrastructure.

Quantifiable outcome

• 6.8x year-on-year increase in worker wages
• +6 more outcomes

Named customers10 records

Segments4 records

Ideal customer profiles4 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability9 records

Feature4 records

Partnerships

14 partnerships are on record, tiered core and supporting.

• AnthropiccoreStrategic or Co-development Partner · 1 February 2026Collaboration with Anthropic to deploy AI models in digital infrastructure, education, agricultural productivity, and healthcare across the Global South. Anthropic has entered partnerships with Indian nonprofits including Karya to apply AI models to various sectors and evaluate model performance across Indian languages.
• Microsoft ResearchcoreStrategic or Co-development PartnerCollaboration on Samiksha evaluation framework and various data collection projects including conversational speech data and LLM evaluations.
• GooglecoreStrategic or Co-development PartnerPartnership for Project Vaani (speech data collection) and translation projects covering 16 long-tail languages with 1.6M+ words translated.
• OpenAIcoreStrategic or Co-development PartnerPartnership for AI data collection and joint delivery of Nonprofit AI Jam sessions in India across Bengaluru, Mumbai, Kolkata, and New Delhi.
• Mozilla FoundationsupportingStrategic or Co-development PartnerPartnership established in August 2025 to support work across conversational data.
• Digital GreensupportingChannel Partner/ Reseller/ DistributorPlatform deployment partner in Kenya for deploying ethical digital work using Karya's platform technology.
• PrathamcoreStrategic or Co-development PartnerCollaboration on educational initiatives and AI tools for learning in India.
• Central Square FoundationsupportingStrategic or Co-development PartnerPartnership on education initiatives and public sector AI collaborations.
• EkStep FoundationcoreStrategic or Co-development PartnerNon-profit partner in the Astitva Prakalp project for Dehwali Bhili language AI model development.
• AI4Bharat at IIT MadrascoreStrategic or Co-development PartnerResearch partner in the Astitva Prakalp project providing technical expertise for language AI model development.
• Karnataka Health Promotion Trust (KHPT)coreStrategic or Co-development PartnerBIRAC and Grand Challenges India-supported project for adolescent health datasets and technology development.
• Government of IndiacoreStrategic or Co-development PartnerMultiple collaborations to build and open source critical contextual datasets and digital public goods. Partner on Samiksha evaluation framework and various AI initiatives.
• Aapti InstitutesupportingStrategic or Co-development PartnerPartnership to develop Code of Conduct for ethical data work, formalizing principles around fair pay, informed consent, grievance redressal, and transparency.
• The/Nudge InstitutecoreStrategic or Co-development PartnerPartnership on Samudaye platform and various digital livelihoods initiatives.

Scale indicators13 records

Recent moves9 records

Expansion highlights7 records

Direct peers

• Sama: Provides ethical AI data annotation and model evaluation services, with operations in East Africa and India and a mission-focused positioning on dignified digital work. Closest mission-aligned peer to Karya — both serve frontier AI clients while explicitly employing low-income workers at fair wages.
• Toloka AI: Crowdsourcing platform for AI data labeling across diverse languages and modalities, with strong presence in emerging markets. Directly comparable to Karya's Platform by Karya in offering scalable, distributed data annotation for AI model development.
• Surge AI: Specialized data labeling platform focused on RLHF and nuanced NLP tasks for frontier AI labs. Comparable to Karya in serving frontier model developers with high-quality human evaluation, though Surge operates primarily with skilled contract annotators rather than rural digital workers.
• CloudFactory: Cloud-based workforce provider for AI data labeling, with operations in Nepal and Kenya employing workers from underserved communities. Comparable to Karya in mission-aligned employment model and AI data services for enterprise clients.
• Labelbox: Data labeling and evaluation platform for AI applications, serving enterprises building production ML systems. Comparable to Karya's Platform by Karya in offering annotation tooling and workforce management for AI data pipelines.
• Defined.ai: Provides ethical AI training data including speech, text, and multimodal datasets across many languages, with a focus on fair compensation and transparency. Comparable to Karya in multilingual data offerings and ethical positioning, though operates as a data marketplace rather than a workforce platform.
• iMerit: India-based data annotation and AI services company providing workforce solutions for AI/ML pipelines across computer vision, NLP, and content services. Directly comparable given Indian operational base, focus on skilled workforce development, and enterprise AI customer base.

Emerging players

• AI4Bharat: Open-source language AI research group at IIT Madras focused on Indic-language NLP, including datasets, models, and benchmarks. A research partner to Karya on Astitva Prakalp but a comparable emerging player in the Indic-language AI ecosystem with overlapping speech and NLP dataset capabilities.

Broad incumbents

• Scale AI: One of the largest AI data labeling and evaluation platforms globally, serving frontier AI labs with annotation across text, image, speech, and RLHF. Directly comparable to Karya in customer base (frontier labs) and core offering (high-quality training data), though Scale operates at far broader geographic and language scale.
• Appen: Global leader in training data for AI, providing data annotation, collection, and evaluation across 235+ languages. Comparable to Karya in multilingual speech and text annotation, but operates at vastly larger scale and with a more commercial (vs. mission-focused) model.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks5 records

Key highlights7 records

Customer concentration

• Xtwitter.com/karya_inc
• LinkedInlinkedin.com/company/karya-inc

Revenue estimate

Valuation estimate

Number of profiles

Profiles15 records

Funding overview

Funding rounds2 records

Investors2 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about karya