NLP Cloud
The facts below are what the maker publishes about NLP Cloud, checked and dated. Our editors have not written about it yet.
Plans and prices
From NLP Cloud's own pricing page, checked 21 Sep 2026.
| Plan | Price | What it gives you |
|---|---|---|
| Large Language Models | Priced on request | 50 parallel requests · 50 GPU requests per minute for Whisper · 350 GPU requests per minute for GPT-OSS 120B and LLaMA 3 · $80/day shown on the pricing page |
| Pay-As-You-Go | Free | 10 parallel requests · CPU $0.003/request · GPU $0.005/request · $15 initial credit · asynchronous mode · $0 / month + usage |
| Free | Free | 2 parallel requests · variable rate limits · pre-trained models · $0 |
| Starter | $29 a month | 10 parallel requests · CPU pre-trained models · $29/month ($1/day) |
| Full | $59 a month | 20 parallel requests · CPU pre-trained models · $59/month ($2/day) |
| Starter GPU | $99 a month | 10 parallel requests · GPU rate limits by model · $99/month ($3.5/day) |
| Full GPU | $199 a month | 20 parallel requests · GPU rate limits by model · $199/month ($7/day) |
| Enterprise | $229 a month | 40 parallel requests · CPU pre-trained models · $229/month ($7/day) |
| Basic Fine-tuning | $399 a month | 1 deployed model · 1 dedicated server · 1,024-token context · 5 parallel requests · 3 fine-tunings included monthly · $399/month ($13/day) |
| Enterprise GPU | $699 a month | 40 parallel requests · GPU rate limits by model · $699/month ($23/day) |
| Advanced Fine-tuning | $990 a month | 1 deployed model · 1 dedicated server · 16,384-token context · 20 parallel requests · 3 fine-tunings included monthly · $990/month ($33/day) |
The facts
Each one read from NLP Cloud's own pages, with the date it was checked.
What it does
- API authentication
- API calls require HTTPS and a token supplied in the Authorization header.checked 15 Oct 2026
- API limits
- When a quota or rate limit is reached, the service returns HTTP 429 and advises retrying later.checked 15 Oct 2026
- Data handling
- The company says it cannot see customer data, does not store it and does not use it to train its own models.checked 15 Oct 2026
- Documentation depth
- The documentation covers installation, authentication, versioning, models, fine-tuning, custom-model uploads, GPU use, multilingual processing, asynchronous mode, endpoints, teams, rate limiting and errors.checked 15 Oct 2026
- Feature: asynchronous processing
- The API offers asynchronous requests for large inputs, with results retained for 48 hours before automatic deletion.checked 15 Oct 2026
- Feature: custom models
- Customers can fine-tune supported models or upload their own Hugging Face Transformers and spaCy models for deployment.checked 15 Oct 2026
- Feature: semantic search
- Users can create semantic-search or retrieval-augmented models from their own documentation, contracts or other domain data.checked 15 Oct 2026
- Feature: speech recognition
- Its automatic speech recognition endpoint transcribes audio or video, detects language, adds punctuation and word timestamps, and supports 100 languages.checked 15 Oct 2026
- Feature: translation
- The translation service uses NLLB 200 and advertises translation across 200 languages with automatic input-language detection.checked 15 Oct 2026
- Founder and legal name
- Crunchbase lists Julien Salinas as founder and CTO and gives the legal name as NLP Cloud.checked 15 Oct 2026
- Integrations
- The API documentation lists official client libraries for Python, Ruby, Go, Node.js and PHP.checked 15 Oct 2026
- Limit: GPU pricing
- G2 reviewers repeatedly say GPU plans can become expensive as usage grows and some would prefer lower rates.checked 15 Oct 2026
- Limit: model quality and latency
- G2 reviews mention slower responses from some large transformer endpoints and occasional output inaccuracies or weaker domain-specific accuracy.checked 15 Oct 2026
- Limit: training workflow
- G2 reviewers note that some custom models must be trained outside the platform and then uploaded because in-platform training options are limited.checked 15 Oct 2026
- Notable news: multilingual add-on
- A 2023 LinkedIn announcement said NLP Cloud made its models available in 200 languages without an additional multilingual charge.checked 15 Oct 2026
- On-premise deployment
- Most models can run on a customer's isolated servers without an internet connection, with engineers available to assist deployment.checked 15 Oct 2026
- Praise: API usability
- G2 reviewers repeatedly praise the clean, simple API and easy integration for deploying NLP models without managing serving infrastructure.checked 15 Oct 2026
- Praise: speed and reliability
- G2 reviewers commonly describe inference as fast or reliable enough for real-time and production workloads.checked 15 Oct 2026
- Praise: support
- Several G2 reviewers describe NLP Cloud support as responsive and helpful.checked 15 Oct 2026
The company
- Company headquarters
- LinkedIn lists headquarters at 447 Broadway Street, New York, NY 10013, United States.checked 15 Oct 2026
- Company size
- LinkedIn lists NLP Cloud as a self-owned software-development company with 11–50 employees.checked 15 Oct 2026
- Founded
- LinkedIn states that NLP Cloud was founded in 2020.checked 15 Oct 2026
- Funding and ownership
- Crunchbase describes NLP Cloud as private and lists no disclosed funding amount; a 2022 interview characterized the business as bootstrapped.checked 15 Oct 2026
- Headquarters
- 447 Broadway, New York City, United Stateschecked 23 Sep 2026
