UNPKG

mcp-omnisearch

Version:

MCP server for integrating Omnisearch with LLMs

56 lines (43 loc) • 4.65 kB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
# Provider selection MCP Omnisearch registers only providers with configured API keys. Use the smallest provider that matches the job, then switch providers when you need a different ranking style, operator model, or processing mode. ## Search providers | Provider | API key | Best for | Operators and filters | | ----------------- | ---------------- | --------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------- | | `brave` | `BRAVE_API_KEY` | Privacy-oriented web search, native search operators, exact discovery | Passes rich operators through in the query string and merges `include_domains` / `exclude_domains` into `site:` / `-site:` clauses. | | `kagi` | `KAGI_API_KEY` | High-quality web results and focused research | Preserves query operators, uses Kagi request parameters for `filetype:` and `before:` / `after:` dates. | | `tavily` | `TAVILY_API_KEY` | Factual/cited search and API-native filtering | Translates supported operators into Tavily fields: domains, dates, exact phrases, and country. | | `exa` | `EXA_API_KEY` | Semantic/neural search and discovery | Supports domain filters through request parameters; optimized for meaning rather than exact operator syntax. | | `kagi_enrichment` | `KAGI_API_KEY` | Specialized Kagi enrichment indexes | Use when enrichment/specialized-index results are desired rather than general web results. | ## AI answer providers | Provider | API key | Best for | | -------------- | ---------------- | ---------------------------------------------- | | `kagi_fastgpt` | `KAGI_API_KEY` | Fast sourced answers. | | `exa_answer` | `EXA_API_KEY` | Semantic AI answers grounded in Exa retrieval. | | `linkup` | `LINKUP_API_KEY` | Deep agentic search with sources. | ## GitHub provider | Provider | API key | Search types | Syntax | | --------------- | ---------------- | ------------------------------- | --------------------------------------------------------------------------------------------- | | `github_search` | `GITHUB_API_KEY` | `code`, `repositories`, `users` | GitHub qualifiers such as `filename:`, `path:`, `repo:`, `user:`, `language:`, and `in:file`. | Use a GitHub personal access token with no scopes selected for public search only. See [troubleshooting](troubleshooting.md#github-token-setup). ## Processing providers | Provider | API key | Modes | Best for | | ----------- | ------------------- | ---------------------------------------------- | -------------------------------------------------------------------------- | | `tavily` | `TAVILY_API_KEY` | `extract` | Content extraction with optional `extract_depth`. | | `kagi` | `KAGI_API_KEY` | `summarize` | Summaries of pages, videos, and podcasts. | | `firecrawl` | `FIRECRAWL_API_KEY` | `scrape`, `crawl`, `map`, `extract`, `actions` | Scraping, crawling, site maps, structured extraction, and browser actions. | | `exa` | `EXA_API_KEY` | `contents`, `similar` | Page content retrieval and semantically similar URLs. | ## Provider choice cheatsheet - Need native operators like `filetype:pdf`, `intitle:`, or `before:`? Start with `brave` or `kagi`. - Need API-level domain/date/country filtering? Use `tavily`. - Need semantic discovery, similar pages, or meaning-based results? Use `exa`. - Need source-grounded narrative answers? Use `ai_search` with `kagi_fastgpt`, `exa_answer`, or `linkup`. - Need to crawl/scrape/map a site? Use `web_extract` with `firecrawl`. - Need public code/repository/user discovery? Use `github_search`.