DeepSeekCost
DeepSeek peak and off-peak routing experiment
A worked cost model for DeepSeek Flash and V4 Pro under the official peak schedule, with a Python routing check and the assumptions that still need human review.
Product updates
Tokuse Changelog
A running record of DeepSeek model support, routing, pricing tools, and developer resources. Each entry links to the page or download it changed.
Integrations
Dify, Open WebUI, and OpenAI Node.js guides
Three new setup guides cover the OpenAI-compatible provider in Dify, the OpenAI connection in Open WebUI, and a Node.js example with the official OpenAI SDK.
Starter kitDeepSeek
DeepSeek starter kit 0.1.0
An MIT-licensed package with a Python client, endpoint checker, cost estimator, tests, and Continue, Cline, LangChain, and Cursor configuration examples.
ToolsDeepSeek
DeepSeek cost calculator with peak and cache inputs
The calculator now models cached input, uncached input, output tokens, and the official peak and off-peak schedule. Results can be shared or embedded with the source link intact.
Integrations
DeepSeek integration center
Copy-ready configurations for Continue, Cline, LangChain, and Cursor, all pointed at one OpenAI-compatible endpoint.
GuidesDeepSeek
Four DeepSeek guides published
Current pricing and cache math, a Chinese cost guide, an OpenAI-compatible setup walkthrough, and a Flash vs V4 Pro routing comparison.
Want to test a model first?
The calculator and integration guides do not require an account. Price a real workload first, then copy the same base URL into your tool.