TokenCalculator.dev
TokenCalculator.dev is an independent, privacy-first browser-based utility for developers, prompt engineers, and teams analyzing prompt size, API input cost, and context-window utilization before calling an LLM. It keeps prompt, document, source-code, data-file, and image content in the browser while exposing the calculation method and linking changing model facts to public sources.
At a Glance
- Developers and AI/ML engineers
- Prompt engineers
- Product and engineering teams building LLM applications
- Data scientists operating document-ingestion or RAG pipelines
- +2 more
AI Tools by TokenCalculator.dev
(1)
TokenCalculator.dev
Local Token Counter for LLMs
Discussions
No discussions yet
Be the first to start a discussion about TokenCalculator.dev
Latest News
TokenCalculator.dev featured in Desifounder's Week 36 project spotlight
TokenCalculator.dev published on Webspot as a privacy-first browser LLM token calculator
Private LLM token counting for enterprise prompts, files, and code
How to calculate tokens in PDFs, DOCX, code, and images
Products & Services
Free browser-based calculator for pasted text, PDF, DOCX, code, structured data, and PNG/JPEG/WebP/GIF images. It returns provider/model token views, context-window fit, and estimated input cost; text and files are processed locally.
Public v1 API at https://tokencalculator.dev/api/v1 with /models, /count, /batch, /compare, and /health endpoints. It accepts text, messages, already-extracted files, image dimensions, output-token assumptions, and cached-input assumptions; the documentation says no API key is required for the initial public service.
Local JavaScript/Node SDK for measuring models and comparing models without sending private prompt content to the TokenCalculator.dev API.
Workflow integration for token counting, cost measurement, and model comparison in self-hosted or cloud n8n workflows.
Market Position
TokenCalculator.dev competes with general-purpose LLM token calculators such as Token Calculator (tokencalculator.com) and Token-Calculator.net. Its differentiation is the combination of browser-local processing, multimodal input support for documents/code/data/images, explicit exact-versus-projected methodology, context-fit checks, and simultaneous OpenAI/Claude/Gemini/DeepSeek comparison. Unlike a simple text counter, it emphasizes measuring the workload that will actually reach a model while clearly warning that provider-side framing and invoices can differ.
Founding Story
The project was created around the problem that the material needing measurement—confidential prompts, internal policies, source code, reports, datasets, and images—is often the material teams are least willing to upload to a separate counting service. Its initial vision was a browser-local preflight tool that measures real multimodal workloads without creating an additional content-disclosure path.
Business Model
Revenue Model
The browser calculator is offered free. The public HTTP API is initially available without an API key and is rate-limited; a local npm SDK and workflow integrations provide an alternative execution boundary for private content. No paid monetization or revenue figure is stated in the reviewed sources.
Pricing Tiers
Browser-based token counting and cost estimation for text, files, and images.
No API key required for the initial public service; documented limits include 60 requests per minute per route/IP/Cloudflare edge, 256 KiB JSON bodies, 200 KiB combined text/message/file content, batches up to 25 measurements, and comparisons across up to 10 models.
Target Markets
- Developers and AI/ML engineers
- Prompt engineers
- Product and engineering teams building LLM applications
- Data scientists operating document-ingestion or RAG pipelines
- Teams handling confidential or regulated prompts and source material
- Automation users of n8n and Activepieces
- Preflight confidential prompts and agent/system instructions before an API call
- Budget document-ingestion and RAG workloads using PDF, DOCX, source, and data files
- Plan multimodal vision requests using image dimensions and provider formulas
- Measure repository-review payloads containing code, configuration, and documentation
- Compare model context fit and input economics before selecting a provider
- Estimate batch-processing and recurring daily/monthly workload costs