# BrandLabs

> BrandLabs builds private, safe AI for health and wellness companies and proves it with evaluation. Offices in New York and Mumbai. Works with health and wellness companies worldwide, from seed stage to Series B. Contact: contact@brandlabs.app, WhatsApp +91 95947 02015

Key facts:
- Wellness AI Safety Check (2–4 weeks · Fixed scope): We test your health or wellness chatbot against US state chatbot laws and clinical-safety basics, then fix what fails.
- Private-by-design sprint (4–8 weeks · Fixed scope): Engineering for India's DPDP Act and US privacy rules: consent, redaction, breach workflows and, if you want them, private models.
- Forward-deployed AI pod (Monthly · Ongoing): A senior pod of four to five joins your team and ships evaluated AI features every week.
- Private AI lab: BrandLabs fine-tunes and runs open medical models on NVIDIA DGX Spark machines it owns. Only de-identified or synthetic data reaches them, and no third-party AI API is used. US patient data stays in the client's cloud under a BAA.
- Free DPDP scanner: a skill for Claude Code and Cursor that checks a codebase against India's DPDP Act 2023 and DPDP Rules 2025, listing each gap with file, line and section. Install: `npx skills add brandlabs/dpdp-scan`.
- Precision (Live): Score every AI answer against your own rubric: a 0–10 accuracy score, written analysis and flagged phrases, one at a time or in bulk.
- Correct (Beta): Every compliance, regulation and tax deadline for an Indian business in one calendar, with alerts and an AI assistant for regulatory questions.
- Production results: Myna Bolo (WhatsApp women's health assistant, 100,000+ conversations, 23,700+ users), Myna Voice (5,269 calls), M-Health (39,334 downloads).
- BrandLabs reports describe tests performed. BrandLabs does not certify legal compliance and is not a law firm.

## Services
- [Services](https://www.brandlabs.app/services): Safety Check, Private-by-design sprint, forward-deployed AI pod
- [Free DPDP scanner](https://www.brandlabs.app/dpdp-scanner): What it checks, install steps, FAQ
- [Private AI](https://www.brandlabs.app/private-ai): On-device de-identification, private model training on NVIDIA DGX Spark, comparison with cloud medical NLP APIs

## Case studies
- [Myna Bolo](https://www.brandlabs.app/case-studies/myna-bolo): A multilingual WhatsApp companion for sexual and reproductive health, built for women with low literacy in rural India.
- [Myna Voice](https://www.brandlabs.app/case-studies/myna-voice-ai): Low-latency phone AI for women who can't read or type.
- [M-Health](https://www.brandlabs.app/case-studies/m-health): A trusted path to reproductive care, in local languages.
- [Qi Health](https://www.brandlabs.app/case-studies/qi-health): Telehealth and custom herbal care plans for women's health.
- [SatvikScan](https://www.brandlabs.app/case-studies/satvikscan): An in-aisle label scanner that gives a clear dietary verdict in seconds.

## Products
- [Precision](https://www.brandlabs.app/our-lab/precision): LLM evaluation platform
- [Correct](https://www.brandlabs.app/our-lab/correct): Compliance management for Indian businesses

## Learn: health AI library
Sourced, dated reference pages on health AI regulation, platforms, open source, research, open data and the market. Plain Markdown: add .md to any URL.
- [Cloud medical NLP APIs vs on-device de-identification](https://www.brandlabs.app/learn/cloud-medical-nlp-vs-on-device-deidentification) (Comparison, reviewed 2026-11-14): AWS Comprehend Medical, Azure Text Analytics for health and Google's Healthcare Natural Language API compared with open-source tools that run on your own hardware.
- [How to build crisis detection into a health chatbot](https://www.brandlabs.app/learn/crisis-detection-in-health-chatbots) (Guide, reviewed 2026-11-14): A practical guide to detecting suicide and self-harm risk in health and wellness chatbots, routing to helplines by country, and testing that it works.
- [How to de-identify health chatbot logs](https://www.brandlabs.app/learn/de-identifying-chat-logs) (Guide, reviewed 2026-11-14): A step-by-step guide to removing names, phone numbers and IDs from health chatbot transcripts before they reach logs, analytics or model training.
- [Digital health funding in 2026: where the money is going](https://www.brandlabs.app/learn/digital-health-funding-2026) (Market, reviewed 2026-11-14): US digital health startups raised $11.1B in the first nine months of 2026, per Rock Health. What the numbers say about AI and what buyers now expect.
- [The EU AI Act for health AI: what applies now and later](https://www.brandlabs.app/learn/eu-ai-act-health-ai) (Regulation, reviewed 2026-11-14): Article 50 chatbot transparency since 2 August 2026, high-risk duties moved to December 2027 and August 2028, and how medical devices fit in.
- [Health AI briefing: November 2026](https://www.brandlabs.app/learn/health-ai-briefing-november-2026) (Briefing, reviewed 2026-11-14): Oura's $625M deal for Counsel Health, new rounds for Charta, Zus, Alterra Labs and Akara, the EU AI Act delay in force, and the US and India deadlines ahead.
- [Health AI regulation around the world: a 2026 tracker](https://www.brandlabs.app/learn/health-ai-regulation-around-the-world) (Regulation, reviewed 2026-11-14): The rules that apply to health and wellness AI in the US, EU, UK, India, Saudi Arabia, the UAE and Singapore, with the dates that matter in 2026 and 2027.
- [Building a health chatbot: 12 do's and don'ts](https://www.brandlabs.app/learn/health-chatbot-dos-and-donts) (Do's and don'ts, reviewed 2026-11-14): Practical rules for teams shipping a health or wellness chatbot: disclosure, crisis handling, minors, medical claims, privacy and testing.
- [Health data laws compared: HIPAA, GDPR, DPDP, PDPL and the UAE](https://www.brandlabs.app/learn/health-data-laws-compared) (Comparison, reviewed 2026-11-14): Who each health data law covers, consent, cross-border transfers, breach reporting and penalties in the US, EU, India, Saudi Arabia and the UAE.
- [HealthBench explained: what OpenAI's health benchmark measures](https://www.brandlabs.app/learn/healthbench-explained) (Research, reviewed 2026-11-14): A plain-English breakdown of HealthBench: 5,000 conversations, 262 physicians, 48,562 rubric criteria, and what it does and does not tell you about a health chatbot.
- [India's DPDP Rules 2025 for health apps: duties and dates](https://www.brandlabs.app/learn/india-dpdp-rules-health-apps) (Regulation, reviewed 2026-11-14): What the DPDP Rules 2025 require of health and wellness apps: notice, consent, children's data, security, breaches and the 13 May 2027 deadline.
- [Speech AI for Indian languages in health voice assistants](https://www.brandlabs.app/learn/indian-language-speech-ai) (Comparison, reviewed 2026-11-14): Sarvam, AI4Bharat, Bhashini and global providers compared for Hindi, Marathi, Tamil and other Indian languages, and how to test speech models on your own callers.
- [Indian public health data sources for AI and research](https://www.brandlabs.app/learn/indian-public-health-data-sources) (Open data, reviewed 2026-11-14): HMIS, NFHS-5, the Indian Food Composition Tables, the ABDM sandbox and data.gov.in: what each offers, how to access it, and its limits.
- [LLM evaluation tools compared for health AI teams](https://www.brandlabs.app/learn/llm-evaluation-tools-compared) (Comparison, reviewed 2026-11-14): Promptfoo, DeepEval, Inspect, Langfuse, Arize Phoenix and Braintrust: licences, strengths and which fits a regulated health product.
- [What 2026 research says about chatbots in mental health crises](https://www.brandlabs.app/learn/mental-health-chatbot-safety-research-2026) (Research, reviewed 2026-11-14): A plain-English summary of 2026 studies on how AI chatbots handle suicide risk and mental health crises, and what product teams should take from them.
- [Open health datasets for AI, and what each is good for](https://www.brandlabs.app/learn/open-health-datasets-for-ai) (Open data, reviewed 2026-11-14): MIMIC-IV, Synthea, MedQA, MedMCQA, PubMedQA, HealthBench, NHANES, India's NFHS-5 and WHO data: access rules, sizes and the best use for each.
- [Open medical language models in 2026 and their licences](https://www.brandlabs.app/learn/open-medical-llms-and-licences) (Open source, reviewed 2026-11-14): MedGemma, Meditron, BioMistral and OpenBioLLM compared: sizes, published MedQA results and licence terms that matter before you build a product.
- [Open-source healthcare AI projects worth knowing in 2026](https://www.brandlabs.app/learn/open-source-healthcare-ai-projects) (Open source, reviewed 2026-11-14): Ten open projects for clinical NLP, de-identification, medical imaging, synthetic data, FHIR and medical language models, and what each one is good for.
- [Saudi Arabia's PDPL and health data: a guide for AI teams](https://www.brandlabs.app/learn/saudi-pdpl-health-data) (Regulation, reviewed 2026-11-14): How Saudi Arabia's Personal Data Protection Law treats health data: explicit consent, transfer rules, SDAIA enforcement since September 2024 and penalties.
- [Singapore's rules for health AI: AIHGle 2.0, HSA and the PDPA](https://www.brandlabs.app/learn/singapore-health-ai-rules) (Regulation, reviewed 2026-11-14): Singapore's refreshed AI in Healthcare Guidelines (March 2026), HSA's software medical device guidance and PDPA duties, explained for health AI teams.
- [UAE health data law: localisation rules for health tech](https://www.brandlabs.app/learn/uae-health-data-law) (Regulation, reviewed 2026-11-14): Federal Law No. 2 of 2019 keeps UAE health data in the country by default. Here are the exceptions, the penalties and what it means for AI products.
- [UK health AI regulation: MHRA, AI Airlock and UK GDPR](https://www.brandlabs.app/learn/uk-health-ai-regulation) (Regulation, reviewed 2026-11-14): How the UK regulates health AI in 2026: AI as a medical device under the MHRA, the AI Airlock sandbox now in phase 3, and UK GDPR for health data.
- [US state AI chatbot laws: what health and wellness apps must do](https://www.brandlabs.app/learn/us-state-ai-chatbot-laws) (Regulation, reviewed 2026-11-14): California SB 243, New York's companion law, Washington HB 2225, Oregon, Tennessee and more: disclosure, crisis protocols, minors and who can sue.
- [WhatsApp Business Platform for health apps: pricing and providers](https://www.brandlabs.app/learn/whatsapp-business-platform-for-health) (Comparison, reviewed 2026-11-14): How Meta's per-message pricing works since July 2025, free service windows, choosing a provider, and the privacy questions a health chatbot on WhatsApp must answer.
- [Women's health AI startups to watch in 2026](https://www.brandlabs.app/learn/womens-health-ai-startups-2026) (Startups, reviewed 2026-11-14): AI companies in women's health that raised in 2026, from prenatal ultrasound to breast screening and cycle tracking, plus where femtech funding stands.

## Insights
- [PHI minimization in RAG pipelines](https://www.brandlabs.app/blog/phi-minimization-rag): Chunking, redaction, retrieval filters, and eval hooks so regulated AI answers without stuffing patient data into every context window.
- [DPDP consent UX that founders can actually ship](https://www.brandlabs.app/blog/dpdp-consent-ux): Practical notice, purpose limitation, and consent patterns for AI products selling to Indian regulated buyers — without freezing the roadmap.
- [BAA before PHI: the sequencing most teams get wrong](https://www.brandlabs.app/blog/baa-before-phi): A decision tree for LLM vendors, subprocessors, and when synthetic data is enough — before anyone pastes real patient data into a prompt.
- [HIPAA vs DPDP: one product, two postures](https://www.brandlabs.app/blog/hipaa-vs-dpdp-posture): Shared primitives for regulated AI — access control, logs, minimization — and the jurisdiction-specific deltas US and Indian buyers actually diligence.
- [Choosing an LLM vendor under regulated constraints](https://www.brandlabs.app/blog/model-vendor-risk): BAA availability, retention defaults, region options, and eval portability — how Brandlabs shortlists model providers for healthcare and regulated buyers.
- [Audit logging for every inference call](https://www.brandlabs.app/blog/audit-logs-for-llm-calls): What to log for HIPAA-aware LLM systems, what never to log, how long to keep it, and who can open the trail.
- [Safety evals for clinical and health chatbots](https://www.brandlabs.app/blog/safety-evals-clinical): How to evaluate refusal, escalation, medical-advice boundaries, and hallucinated dosages — before a health assistant meets real users.
- [Building a prompt regression dataset that ages well](https://www.brandlabs.app/blog/prompt-regression-datasets): Golden sets, adversarial sets, and ownership so your LLM suite still means something after six months of prompt thrash.
- [Three eval dimensions: functional, qualitative, safety](https://www.brandlabs.app/blog/eval-dimensions-functional-qualitative-safety): The Brandlabs Discovery Sprint pattern for setting LLM eval targets in weeks 1–3 — before feature thrash eats the timeline.
- [CI that blocks deploys on failing evals](https://www.brandlabs.app/blog/ci-evals-every-commit): Sampling, cost caps, flake handling, and which LLM eval suites must stay required vs nightly — so production cannot silently regress.
- [Who sits in an embedded AI pod](https://www.brandlabs.app/blog/embedded-pod-composition): FDE, senior engineers, product design, and compliance-as-needed — roles and anti-patterns for shipping regulated AI in 12–20 weeks.
- [How US healthcare startups should buy AI build partners](https://www.brandlabs.app/blog/us-healthcare-startup-ai-buy): Diligence questions on BAA fluency, eval discipline, and clinical workflow empathy — before you hire a studio to ship your AI product.
- [The two-week Discovery Sprint, unbundled](https://www.brandlabs.app/blog/discovery-sprint-two-weeks): Inputs, outputs, kill criteria, and how Brandlabs maps a Discovery Sprint into a 12–20 week zero-to-PMF engagement.
- [Regulated-industry AI buyer’s guide (India)](https://www.brandlabs.app/blog/regulated-ai-buyer-guide-india): Questions CFOs and compliance officers should ask before signing an AI vendor — DPDP, updates, liability, and evidence — not just a demo.
- [Zero to PMF in 12–20 weeks: the phase model](https://www.brandlabs.app/blog/twelve-to-twenty-week-pmf): Foundation → build to alpha → production-harden → PMF hunt — with eval and compliance gates at each phase.
- [Trust UX for telehealth and women’s health platforms](https://www.brandlabs.app/blog/telehealth-trust-ux): Privacy, language, and stigma-aware product patterns from shipping telehealth for women’s health — so care feels safe enough to start.
- [Safety engines for WhatsApp health assistants](https://www.brandlabs.app/blog/whatsapp-health-bots-safety): Escalation, localization, and content safety patterns for healthcare chat on WhatsApp — from shipping a women’s health assistant in production.
- [Voice AI for low-literacy health access](https://www.brandlabs.app/blog/voice-ai-low-literacy-health): Latency, vernacular STT/TTS, and when voice beats chat — lessons from rebuilding a voice assistant for rural women’s health support.
- [AI for Indian MSME compliance (lessons from Correct)](https://www.brandlabs.app/blog/india-msme-compliance-ai): Productizing 69,000+ obligations and regulations that change ~15 times a day — without turning an AI assistant into unauthorized legal advice.
- [Evals are the product](https://www.brandlabs.app/blog/evals-are-the-product): Why we treat the eval suite as a first-class deliverable — and how it changes how we negotiate scope.
- [Shipping on Fridays without breaking trust](https://www.brandlabs.app/blog/shipping-on-fridays): A pattern for weekly demos that are evidence, not theatre. What we ship, what we don't, and what we say out loud.
- [Compliance as architecture, not paperwork](https://www.brandlabs.app/blog/compliance-as-architecture): HIPAA and DPDP shouldn't be a last-mile checklist. Here's how we bake them into the first commit.
- [Every postmortem becomes a regression test](https://www.brandlabs.app/blog/postmortems-as-regressions): The cultural and technical loop that turns production incidents into permanent guardrails.
- [What a forward-deployed engineer actually does](https://www.brandlabs.app/blog/forward-deployed-engineering): Not a project manager, not a consultant. A senior engineer who lives in your Slack and writes the hardest code.

## Optional
- [Topics](https://www.brandlabs.app/blog/topics): Regulated AI architecture, Evals & production AI quality, Healthcare & India, Forward-deployed delivery
- [Careers](https://www.brandlabs.app/careers): Open roles
- [Full text for LLMs](https://www.brandlabs.app/llms-full.txt)
- [RSS](https://www.brandlabs.app/blog/rss.xml)