Nyckel · Feb 2025
Nyckel says it uses transfer-learning-based AutoML on AWS Lambda, can reach high accuracy with as few as 13 samples per class, and relies on massive parallelization to train fast with lower annotation needs.
Public company, workplace, funding, and market signals
Updated Jul 30, 2026
Nyckel is a developer-focused AI platform for reliable predictions, offering managed APIs that let users train custom classification, detection, and search models from examples across image, text, tabular, and multimodal inputs.
Primary product
Managed ML classification/detection APIs for custom image, text, structured-data, and multimodal predictions
Founded
2021
Headquarters
Bay Area, California, United States (LinkedIn); some databases also list San Diego, CA 92109, United States.
Team size
1-10 employees
Industry
Software Development
Sub-industry
AI / machine learning classification APIs and AutoML
Offices
0 jobs at Nyckel
Check back later for new openings
Business model
Stage
Pre-Seed
Total raised
$500K
Latest round
Pre-Seed · Mar 2022
Latest amount
$500K
Mar 2022 · Y Combinator
Investors
Technical, developer-centric, and speed-oriented; the company emphasizes rapid model training, reliable predictions, continuous improvement, and compliance/security (SOC2/HIPAA).
Compensation
No public employee compensation information found.
Pricing
Usage-based pricing with Free, Starter ($149/month), Business ($599/month), and Enterprise/custom tiers; metered by invokes and samples.
Differentiators
Technology
Customers
Competitors
Estimated revenue
Public estimates conflict, roughly $320K/year to about $3M ARR.
Estimated monthly visits
206.9K
Traffic estimate as of May 2026
Nyckel · Feb 2025
Nyckel says it uses transfer-learning-based AutoML on AWS Lambda, can reach high accuracy with as few as 13 samples per class, and relies on massive parallelization to train fast with lower annotation needs.
Nyckel · Aug 2024
Nyckel announced a partnership with Contextual to embed its classification APIs into automated AI workflows for lead prioritization, support-ticket handling, and service work-order classification.
Nyckel · Jun 2024
A customer case study claiming 10–20 samples per class, ~30-second model training/deploy, and savings of hundreds of thousands of dollars versus building in-house.
Nyckel · May 2024
Nyckel launched multimodal classification, letting one function use both image and text inputs together for richer classification tasks.
Nyckel · Jan 2024
Nyckel announced pretrained moderation classifiers for NSFW content, pornography, weapons, offensive language, smoking/alcohol, and image-quality checks.