prime-deepseek-search1.2.0
Web search and page fetch for Prime Agent using DeepSeek, as a kernel-callable Python skill.
Untested on omp
This is a pi package. No one has reported whether it works on omp yet. If you try it, add your result below.
Reports 0
No reports yet.
Readme
prime-deepseek-search
prime-agent package install npm:prime-deepseek-search
Search the web and read pages from Prime Agent using DeepSeek's Search & Fetch APIs. Uses the DeepSeek key saved through /login or DEEPSEEK_API_KEY, with any active model.
print(await deepseek_search("latest OpenUSD security advisories"))
print(await deepseek_search.fetch("https://openusd.org/"))
Search accepts optional allowed_domains or blocked_domains. The existing num_results option now controls the result count (1–50, default 10). Both calls support max_output and timeout; PRIME_AGENT_WEBSEARCH_TIMEOUT sets the default HTTP timeout in seconds.
To use model-mediated server search, set DEEPSEEK_SEARCH_BACKEND=server before starting Prime Agent. DEEPSEEK_SEARCH_MODEL or the existing model keyword selects its DeepSeek model (default deepseek-flash). Fetch remains standalone.
Pricing
Standalone Search and Fetch cost $0.75 per 1,000 successful requests. Failed requests and empty results are free. Returned costs are estimates.
Server search adds model-token charges and can make up to five native searches per call. DeepSeek's temporary waiver for server-side request fees ends November 1, 2026 at 00:00 Beijing time; standard request pricing applies after that. Even during the waiver, model-token costs can make server search more expensive than the standalone API.
Coexists with Prime Agent's bundled Serper websearch skill. Pi users: install pi-deepseek-search instead.