DeepSeek opens Search, Fetch, OCR, ASR and TTS as standalone APIs, with tool fees billed on top of tokens
DeepSeek has made the capabilities behind its app available as standalone APIs: Search and Fetch, OCR, ASR and TTS. It also open-sourced chat-demo, a static web client that connects to the API from the browser. Tool use is billed on top of tokens.
DeepSeek has opened the capabilities behind its app as standalone APIs: Search and Fetch, OCR, ASR for speech recognition and TTS for speech synthesis. Tool use is billed on top of tokens, according to DeepSeek's pricing page.
chat-demo client
To demonstrate the APIs, DeepSeek open-sourced chat-demo, a static web client that connects straight from the browser to the API. It supports deep thinking, search, link reading, attachments and voice input. It can also import a user's DeepSeek chat history.
Tool pricing
Per DeepSeek's pricing page:
| Tool | Price |
|---|---|
| Search (text or images) | $0.75 per 1,000 calls (5 yuan on the China site) |
| Web fetch | $0.75 per 1,000 calls (5 yuan on the China site) |
| File parsing and OCR | $0.225 per 1,000 pages |
| Text-to-speech | $0.45 per 10,000 characters |
| Speech recognition | $0.116 per hour of audio |
Both text-to-speech and speech recognition are currently half price. Search and fetch fees apply whether developers call those APIs directly or the model uses them through its tools.
Relationship to earlier release
The release extends the standalone Search and Fetch APIs that DeepSeek opened on October 17, which run on its in-house search engine.
Token prices for deepseek-flash are unchanged: $0.30 per million cache-miss input tokens and $1.20 per million output tokens at peak hours, with off-peak rates at half those levels.
Sources: the official DeepSeek API change log, the official DeepSeek models and pricing page, the official chat-demo repository, and an IT Home report, as summarized by Pandaily.
What this means
Developers can now call search, page fetching, document parsing, OCR and voice as separate services rather than only through the DeepSeek app. The chat-demo repository provides a reference client that uses these features together. Because search and fetch fees apply whether the APIs are called directly or the model invokes them as tools, any application that enables the model's search tools will incur those per-call charges in addition to token costs.
Related Articles
Analysis finds Blackwell fingerprints in six open-weight models from four Chinese labs
Unemployment Labs reports that six models released since April 2026 by DeepSeek, Moonshot AI, MiniMax and Ant Group carry numerical patterns matching NVIDIA Blackwell GPUs. Earlier models from the same labs match the older Hopper generation. The analysis says it does not establish where the chips were located or whether export controls were violated.
Apple adds $49.99 Intelligence Usage Packs to Logic Pro for extended Apple Intelligence access
Apple has updated Logic Pro with a "Buy More Usage" option, a one-time $49.99 purchase that extends access to Apple Intelligence features with usage limits. The packs cover features that rely on cloud-based models from OpenAI or Google Gemini. Apple is expected to extend the option to other Creator Studio apps this month.
MarkTechPost Compares Token Routing Across 10 Open MoE Models, Including DeepSeek, Qwen, Mixtral, Llama 4 and gpt-oss
MarkTechPost has published a config-verified comparison of how 10 open mixture-of-experts models route tokens. The title names DeepSeek, Qwen, Mixtral, Llama 4 and gpt-oss among the models compared.
DeepSeek Harness Reaches 1.54M Daily Users, Plans Official Plugin Marketplace
DeepSeek Harness, the company's open-source agent runtime, had 1.54 million daily active users as of November 4 and consumes about 92 trillion tokens a day. The team plans an official plugin marketplace and a support program for plugin authors, with no launch dates given.
Comments
Loading...