{ "type": "comparison", "id": "web-search-apis", "title": "Web search APIs for agents", "summary": "Tavily, Exa, Brave Search, Parallel and Perplexity compared on endpoints, index, list price per 1,000 searches, free allowance and MCP support, for agents that need fresh web results.", "author": "Agentica Author", "tags": [ "web-search", "apis", "mcp", "research" ], "published": "2026-10-02", "last_verified": "2026-10-02", "entries": [ "tavily-api", "exa-api", "brave-search-api", "parallel-api", "perplexity-api", "tavily-mcp", "exa-mcp-server", "brave-search-mcp-server", "firecrawl-api", "jina-reader" ], "links": { "html": "https://indexagentica.com/compare/web-search-apis/", "markdown": "https://indexagentica.com/compare/web-search-apis.md", "json": "https://indexagentica.com/api/longform/compare/web-search-apis.json", "source": "https://github.com/Drudley/indexagentica/blob/main/content-long/compare/web-search-apis.md" }, "status": "published", "comparison": { "subjects": [ { "id": "tavily-api", "name": "Tavily API", "summary": "Real-time web layer for AI agents: search, extract, crawl, map and cited research returned as clean LLM-ready content.", "url": "https://indexagentica.com/entries/tavily-api/", "json": "https://indexagentica.com/api/entries/tavily-api.json" }, { "id": "exa-api", "name": "Exa API", "summary": "Search and retrieval API built for AI agents, backed by Exa's own continuously updated web index and search models.", "url": "https://indexagentica.com/entries/exa-api/", "json": "https://indexagentica.com/api/entries/exa-api.json" }, { "id": "brave-search-api", "name": "Brave Search API", "summary": "Web search API on Brave's independent index of 40+ billion pages, with specialized endpoints for search, AI grounding and more.", "url": "https://indexagentica.com/entries/brave-search-api/", "json": "https://indexagentica.com/api/entries/brave-search-api.json" }, { "id": "parallel-api", "name": "Parallel Web APIs", "summary": "Web APIs for AI agents (Search, Extract, Task/Deep Research, FindAll, Monitor) for research, extraction and continuous monitoring.", "url": "https://indexagentica.com/entries/parallel-api/", "json": "https://indexagentica.com/api/entries/parallel-api.json" }, { "id": "perplexity-api", "name": "Perplexity API Platform", "summary": "Perplexity's API platform with Agent, Search and Embeddings APIs for real-time, web-grounded research and Q&A.", "url": "https://indexagentica.com/entries/perplexity-api/", "json": "https://indexagentica.com/api/entries/perplexity-api.json" } ], "criteria": [ "Endpoints", "Index", "Search price", "Free allowance", "MCP server", "Agent extras" ], "rows": { "tavily-api": { "Endpoints": "Search, Extract, Crawl, Map, Research", "Index": "Not stated in the sources used", "Search price": "Credit-based; credits per call depend on the endpoint and depth", "Free allowance": "1,000 API credits per month, no card; free for students", "MCP server": "Yes, remote at mcp.tavily.com/mcp with OAuth or an API key (tavily-mcp)", "Agent extras": "tvly CLI that also installs Agent Skills; docs MCP server" }, "exa-api": { "Endpoints": "Search (several depths), Contents, Answer, Monitors, Exa Agent", "Index": "Own web index", "Search price": "From $4 per 1,000 requests (up to 10 results); contents $1 per 1,000 pages", "Free allowance": "$10 of credits per month plus a $10 onboarding bonus, no payment method", "MCP server": "Yes, hosted at mcp.exa.ai/mcp, no API key needed to start; open source (exa-mcp-server)", "Agent extras": "Agent Skills (npx skills add exa-labs/agent-skills), Claude connector" }, "brave-search-api": { "Endpoints": "Web, news, images, videos, local and LLM context; Answers; spellcheck and autosuggest", "Index": "Own independent index, 40+ billion pages", "Search price": "$5 per 1,000 requests; Answers $4 per 1,000 queries plus tokens", "Free allowance": "$5 of credits every month", "MCP server": "Yes, brave-search-mcp-server (stdio by default, HTTP optional)", "Agent extras": "Goggles to re-rank or exclude domains" }, "parallel-api": { "Endpoints": "Search, Extract, Task (deep research), Responses, Monitor, FindAll", "Index": "Not stated in the sources used", "Search price": "$0.001 to $0.005 per request for 10 results ($1 to $5 per 1,000)", "Free allowance": "Up to 5,000 requests per month free; $5 monthly credit and up to $80 at signup", "MCP server": "Yes (Search MCP and Task MCP)", "Agent extras": "Compressed excerpts sized for LLM context; structured outputs with citations and confidence on Task" }, "perplexity-api": { "Endpoints": "Search API (raw results), Sonar and Agent APIs (grounded answers), Embeddings", "Index": "Not stated in the sources used", "Search price": "$5 per 1,000 requests, no token costs", "Free allowance": "Not stated in the sources used", "MCP server": "Yes (Perplexity API MCP server)", "Agent extras": "Multi-provider Agent API for web-grounded answers" } }, "table": [ { "entry": "tavily-api", "name": "Tavily API", "url": "https://indexagentica.com/entries/tavily-api/", "values": { "Endpoints": "Search, Extract, Crawl, Map, Research", "Index": "Not stated in the sources used", "Search price": "Credit-based; credits per call depend on the endpoint and depth", "Free allowance": "1,000 API credits per month, no card; free for students", "MCP server": "Yes, remote at mcp.tavily.com/mcp with OAuth or an API key (tavily-mcp)", "Agent extras": "tvly CLI that also installs Agent Skills; docs MCP server" } }, { "entry": "exa-api", "name": "Exa API", "url": "https://indexagentica.com/entries/exa-api/", "values": { "Endpoints": "Search (several depths), Contents, Answer, Monitors, Exa Agent", "Index": "Own web index", "Search price": "From $4 per 1,000 requests (up to 10 results); contents $1 per 1,000 pages", "Free allowance": "$10 of credits per month plus a $10 onboarding bonus, no payment method", "MCP server": "Yes, hosted at mcp.exa.ai/mcp, no API key needed to start; open source (exa-mcp-server)", "Agent extras": "Agent Skills (npx skills add exa-labs/agent-skills), Claude connector" } }, { "entry": "brave-search-api", "name": "Brave Search API", "url": "https://indexagentica.com/entries/brave-search-api/", "values": { "Endpoints": "Web, news, images, videos, local and LLM context; Answers; spellcheck and autosuggest", "Index": "Own independent index, 40+ billion pages", "Search price": "$5 per 1,000 requests; Answers $4 per 1,000 queries plus tokens", "Free allowance": "$5 of credits every month", "MCP server": "Yes, brave-search-mcp-server (stdio by default, HTTP optional)", "Agent extras": "Goggles to re-rank or exclude domains" } }, { "entry": "parallel-api", "name": "Parallel Web APIs", "url": "https://indexagentica.com/entries/parallel-api/", "values": { "Endpoints": "Search, Extract, Task (deep research), Responses, Monitor, FindAll", "Index": "Not stated in the sources used", "Search price": "$0.001 to $0.005 per request for 10 results ($1 to $5 per 1,000)", "Free allowance": "Up to 5,000 requests per month free; $5 monthly credit and up to $80 at signup", "MCP server": "Yes (Search MCP and Task MCP)", "Agent extras": "Compressed excerpts sized for LLM context; structured outputs with citations and confidence on Task" } }, { "entry": "perplexity-api", "name": "Perplexity API Platform", "url": "https://indexagentica.com/entries/perplexity-api/", "values": { "Endpoints": "Search API (raw results), Sonar and Agent APIs (grounded answers), Embeddings", "Index": "Not stated in the sources used", "Search price": "$5 per 1,000 requests, no token costs", "Free allowance": "Not stated in the sources used", "MCP server": "Yes (Perplexity API MCP server)", "Agent extras": "Multi-provider Agent API for web-grounded answers" } } ], "verdict": "For raw results to feed your own agent, Exa, Brave and Parallel are the most transparent on price, and Brave is the only one stating index size. Tavily is the easiest all-in-one (search, extract, crawl and research behind one key and an OAuth MCP server). Choose Perplexity or Parallel's Task API when you want the provider to do the research and return a cited answer." }, "entries_detail": [ { "id": "tavily-api", "name": "Tavily API", "summary": "Real-time web layer for AI agents: search, extract, crawl, map and cited research returned as clean LLM-ready content.", "url": "https://indexagentica.com/entries/tavily-api/", "json": "https://indexagentica.com/api/entries/tavily-api.json" }, { "id": "exa-api", "name": "Exa API", "summary": "Search and retrieval API built for AI agents, backed by Exa's own continuously updated web index and search models.", "url": "https://indexagentica.com/entries/exa-api/", "json": "https://indexagentica.com/api/entries/exa-api.json" }, { "id": "brave-search-api", "name": "Brave Search API", "summary": "Web search API on Brave's independent index of 40+ billion pages, with specialized endpoints for search, AI grounding and more.", "url": "https://indexagentica.com/entries/brave-search-api/", "json": "https://indexagentica.com/api/entries/brave-search-api.json" }, { "id": "parallel-api", "name": "Parallel Web APIs", "summary": "Web APIs for AI agents (Search, Extract, Task/Deep Research, FindAll, Monitor) for research, extraction and continuous monitoring.", "url": "https://indexagentica.com/entries/parallel-api/", "json": "https://indexagentica.com/api/entries/parallel-api.json" }, { "id": "perplexity-api", "name": "Perplexity API Platform", "summary": "Perplexity's API platform with Agent, Search and Embeddings APIs for real-time, web-grounded research and Q&A.", "url": "https://indexagentica.com/entries/perplexity-api/", "json": "https://indexagentica.com/api/entries/perplexity-api.json" }, { "id": "tavily-mcp", "name": "Tavily MCP Server", "summary": "Tavily's MCP server with real-time web search, extract, map and crawl tools.", "url": "https://indexagentica.com/entries/tavily-mcp/", "json": "https://indexagentica.com/api/entries/tavily-mcp.json" }, { "id": "exa-mcp-server", "name": "Exa MCP Server", "summary": "Official Exa MCP server connecting agents to Exa for web search, content fetching and multi-step research.", "url": "https://indexagentica.com/entries/exa-mcp-server/", "json": "https://indexagentica.com/api/entries/exa-mcp-server.json" }, { "id": "brave-search-mcp-server", "name": "Brave Search MCP Server", "summary": "Official MCP server for the Brave Search API: web, local, image, video and news search plus LLM context and summarization (stdio or HTTP).", "url": "https://indexagentica.com/entries/brave-search-mcp-server/", "json": "https://indexagentica.com/api/entries/brave-search-mcp-server.json" }, { "id": "firecrawl-api", "name": "Firecrawl API", "summary": "Web data API for AI agents: search the web, scrape any page to clean data and interact with it through one API.", "url": "https://indexagentica.com/entries/firecrawl-api/", "json": "https://indexagentica.com/api/entries/firecrawl-api.json" }, { "id": "jina-reader", "name": "Jina Reader", "summary": "Converts any URL into LLM-friendly Markdown by prefixing it with https://r.jina.ai/.", "url": "https://indexagentica.com/entries/jina-reader/", "json": "https://indexagentica.com/api/entries/jina-reader.json" } ], "related": [ { "type": "stack", "id": "research-agent", "title": "Research agent stack", "url": "https://indexagentica.com/stacks/research-agent/", "json": "https://indexagentica.com/api/longform/stacks/research-agent.json" }, { "type": "guide", "id": "connect-an-agent-to-a-remote-mcp-server", "title": "Connect an agent to a remote MCP server", "url": "https://indexagentica.com/guides/connect-an-agent-to-a-remote-mcp-server/", "json": "https://indexagentica.com/api/longform/guides/connect-an-agent-to-a-remote-mcp-server.json" }, { "type": "stack", "id": "coding-agent-starter", "title": "Coding agent starter stack", "url": "https://indexagentica.com/stacks/coding-agent-starter/", "json": "https://indexagentica.com/api/longform/stacks/coding-agent-starter.json" } ], "sources": [ { "title": "Tavily docs index (llms.txt)", "url": "https://docs.tavily.com/llms.txt", "accessed": "2026-10-02" }, { "title": "Tavily pricing", "url": "https://www.tavily.com/pricing", "accessed": "2026-10-02" }, { "title": "Exa docs index (llms.txt)", "url": "https://exa.ai/docs/llms.txt", "accessed": "2026-10-02" }, { "title": "Exa pricing", "url": "https://exa.ai/pricing", "accessed": "2026-10-02" }, { "title": "Exa docs, Exa MCP", "url": "https://exa.ai/docs/get-started/exa-mcp", "accessed": "2026-10-02" }, { "title": "Brave Search API", "url": "https://brave.com/search/api/", "accessed": "2026-10-02" }, { "title": "Brave Search API pricing", "url": "https://api-dashboard.search.brave.com/documentation/pricing", "accessed": "2026-10-02" }, { "title": "Brave Search MCP server README", "url": "https://github.com/brave/brave-search-mcp-server", "accessed": "2026-10-02" }, { "title": "Parallel pricing", "url": "https://parallel.ai/pricing", "accessed": "2026-10-02" }, { "title": "Parallel docs index (llms.txt)", "url": "https://docs.parallel.ai/llms.txt", "accessed": "2026-10-02" }, { "title": "Perplexity docs, Pricing", "url": "https://docs.perplexity.ai/docs/getting-started/pricing", "accessed": "2026-10-02" }, { "title": "Perplexity docs index (llms.txt)", "url": "https://docs.perplexity.ai/llms.txt", "accessed": "2026-10-02" } ], "front_matter": { "id": "web-search-apis", "type": "comparison", "title": "Web search APIs for agents", "summary": "Tavily, Exa, Brave Search, Parallel and Perplexity compared on endpoints, index, list price per 1,000 searches, free allowance and MCP support, for agents that need fresh web results.", "description": "List prices and free allowances are from each vendor's pricing page on 2026-10-02 and change often; the per-request figures are for a basic search, and deeper modes cost more. Result quality isn't compared here: run your own queries through two or three of these before you choose.", "author": "Agentica Author", "tags": [ "web-search", "apis", "mcp", "research" ], "entries": [ "tavily-api", "exa-api", "brave-search-api", "parallel-api", "perplexity-api", "tavily-mcp", "exa-mcp-server", "brave-search-mcp-server" ], "subjects": [ "tavily-api", "exa-api", "brave-search-api", "parallel-api", "perplexity-api" ], "criteria": [ "Endpoints", "Index", "Search price", "Free allowance", "MCP server", "Agent extras" ], "rows": { "tavily-api": { "Endpoints": "Search, Extract, Crawl, Map, Research", "Index": "Not stated in the sources used", "Search price": "Credit-based; credits per call depend on the endpoint and depth", "Free allowance": "1,000 API credits per month, no card; free for students", "MCP server": "Yes, remote at mcp.tavily.com/mcp with OAuth or an API key (tavily-mcp)", "Agent extras": "tvly CLI that also installs Agent Skills; docs MCP server" }, "exa-api": { "Endpoints": "Search (several depths), Contents, Answer, Monitors, Exa Agent", "Index": "Own web index", "Search price": "From $4 per 1,000 requests (up to 10 results); contents $1 per 1,000 pages", "Free allowance": "$10 of credits per month plus a $10 onboarding bonus, no payment method", "MCP server": "Yes, hosted at mcp.exa.ai/mcp, no API key needed to start; open source (exa-mcp-server)", "Agent extras": "Agent Skills (npx skills add exa-labs/agent-skills), Claude connector" }, "brave-search-api": { "Endpoints": "Web, news, images, videos, local and LLM context; Answers; spellcheck and autosuggest", "Index": "Own independent index, 40+ billion pages", "Search price": "$5 per 1,000 requests; Answers $4 per 1,000 queries plus tokens", "Free allowance": "$5 of credits every month", "MCP server": "Yes, brave-search-mcp-server (stdio by default, HTTP optional)", "Agent extras": "Goggles to re-rank or exclude domains" }, "parallel-api": { "Endpoints": "Search, Extract, Task (deep research), Responses, Monitor, FindAll", "Index": "Not stated in the sources used", "Search price": "$0.001 to $0.005 per request for 10 results ($1 to $5 per 1,000)", "Free allowance": "Up to 5,000 requests per month free; $5 monthly credit and up to $80 at signup", "MCP server": "Yes (Search MCP and Task MCP)", "Agent extras": "Compressed excerpts sized for LLM context; structured outputs with citations and confidence on Task" }, "perplexity-api": { "Endpoints": "Search API (raw results), Sonar and Agent APIs (grounded answers), Embeddings", "Index": "Not stated in the sources used", "Search price": "$5 per 1,000 requests, no token costs", "Free allowance": "Not stated in the sources used", "MCP server": "Yes (Perplexity API MCP server)", "Agent extras": "Multi-provider Agent API for web-grounded answers" } }, "verdict": "For raw results to feed your own agent, Exa, Brave and Parallel are the most transparent on price, and Brave is the only one stating index size. Tavily is the easiest all-in-one (search, extract, crawl and research behind one key and an OAuth MCP server). Choose Perplexity or Parallel's Task API when you want the provider to do the research and return a cited answer.", "sources": [ { "title": "Tavily docs index (llms.txt)", "url": "https://docs.tavily.com/llms.txt", "accessed": "2026-10-02" }, { "title": "Tavily pricing", "url": "https://www.tavily.com/pricing", "accessed": "2026-10-02" }, { "title": "Exa docs index (llms.txt)", "url": "https://exa.ai/docs/llms.txt", "accessed": "2026-10-02" }, { "title": "Exa pricing", "url": "https://exa.ai/pricing", "accessed": "2026-10-02" }, { "title": "Exa docs, Exa MCP", "url": "https://exa.ai/docs/get-started/exa-mcp", "accessed": "2026-10-02" }, { "title": "Brave Search API", "url": "https://brave.com/search/api/", "accessed": "2026-10-02" }, { "title": "Brave Search API pricing", "url": "https://api-dashboard.search.brave.com/documentation/pricing", "accessed": "2026-10-02" }, { "title": "Brave Search MCP server README", "url": "https://github.com/brave/brave-search-mcp-server", "accessed": "2026-10-02" }, { "title": "Parallel pricing", "url": "https://parallel.ai/pricing", "accessed": "2026-10-02" }, { "title": "Parallel docs index (llms.txt)", "url": "https://docs.parallel.ai/llms.txt", "accessed": "2026-10-02" }, { "title": "Perplexity docs, Pricing", "url": "https://docs.perplexity.ai/docs/getting-started/pricing", "accessed": "2026-10-02" }, { "title": "Perplexity docs index (llms.txt)", "url": "https://docs.perplexity.ai/llms.txt", "accessed": "2026-10-02" } ], "related": [ "research-agent", "connect-an-agent-to-a-remote-mcp-server", "coding-agent-starter" ], "last_verified": "2026-10-02", "published": "2026-10-02" }, "markdown": "\n## What these APIs are for\n\nA model's training data is months old, and a general web search engine returns pages built for people. These APIs return results built for a model: clean text or excerpts, sometimes full page contents, in one call, priced per request. All five offer an MCP server, so a harness such as Claude Code or Codex can use them without custom code; see the [remote MCP guide](https://indexagentica.com/guides/connect-an-agent-to-a-remote-mcp-server/).\n\nThey split into two groups:\n\n- **Search primitives** return ranked results and excerpts, and your agent does the reading and reasoning: [Exa](https://indexagentica.com/entries/exa-api/) Search, [Brave Search](https://indexagentica.com/entries/brave-search-api/), [Parallel](https://indexagentica.com/entries/parallel-api/) Search, the [Perplexity](https://indexagentica.com/entries/perplexity-api/) Search API and [Tavily](https://indexagentica.com/entries/tavily-api/) Search.\n- **Research endpoints** run a multi-step search and return a cited answer or a structured result: Tavily Research, Exa Answer and Exa Agent, Parallel Task, Perplexity Sonar and Agent APIs, and Brave Answers. They save agent turns but hide the intermediate steps, which makes them harder to audit.\n\n## Choosing\n\n- **Price per search** ranges from about $1 to $5 per 1,000 basic requests among the vendors that publish per-request prices; deeper modes, more results and page contents add to that. Tavily prices in credits whose cost depends on the endpoint and depth.\n- **Index.** Brave and Exa run their own indexes, and Brave publishes a size (40+ billion pages, with 100+ million page updates a day). The others don't state where results come from in the pages used here.\n- **Reading pages.** If your agent needs full text, check whether contents come in the same call (Exa Contents, Tavily Extract, Parallel Extract) or need a separate reader such as [Firecrawl](https://indexagentica.com/entries/firecrawl-api/) or [Jina Reader](https://indexagentica.com/entries/jina-reader/).\n- **MCP auth.** Exa's hosted server works without a key to start, and Tavily's supports OAuth, which keeps keys out of config files. Brave's server now defaults to stdio and needs `BRAVE_MCP_TRANSPORT=http` for HTTP.\n- **Evaluate on your own queries.** Quality varies by domain (news, code, academic, local). Run a fixed set of 20 to 50 real queries through two or three APIs and compare what your agent does with the results.\n\nFor a full research setup built on these, see the [research agent stack](https://indexagentica.com/stacks/research-agent/).\n", "raw": "---\nid: web-search-apis\ntype: comparison\ntitle: Web search APIs for agents\nsummary: Tavily, Exa, Brave Search, Parallel and Perplexity compared on endpoints, index, list price per 1,000 searches, free allowance and MCP support, for agents that need fresh web results.\ndescription: \"List prices and free allowances are from each vendor's pricing page on 2026-10-02 and change often; the per-request figures are for a basic search, and deeper modes cost more. Result quality isn't compared here: run your own queries through two or three of these before you choose.\"\nauthor: Agentica Author\ntags: [web-search, apis, mcp, research]\nentries: [tavily-api, exa-api, brave-search-api, parallel-api, perplexity-api, tavily-mcp, exa-mcp-server, brave-search-mcp-server]\nsubjects: [tavily-api, exa-api, brave-search-api, parallel-api, perplexity-api]\ncriteria:\n - Endpoints\n - Index\n - Search price\n - Free allowance\n - MCP server\n - Agent extras\nrows:\n tavily-api:\n Endpoints: \"Search, Extract, Crawl, Map, Research\"\n Index: Not stated in the sources used\n Search price: \"Credit-based; credits per call depend on the endpoint and depth\"\n Free allowance: \"1,000 API credits per month, no card; free for students\"\n MCP server: \"Yes, remote at mcp.tavily.com/mcp with OAuth or an API key (tavily-mcp)\"\n Agent extras: \"tvly CLI that also installs Agent Skills; docs MCP server\"\n exa-api:\n Endpoints: \"Search (several depths), Contents, Answer, Monitors, Exa Agent\"\n Index: Own web index\n Search price: \"From $4 per 1,000 requests (up to 10 results); contents $1 per 1,000 pages\"\n Free allowance: \"$10 of credits per month plus a $10 onboarding bonus, no payment method\"\n MCP server: \"Yes, hosted at mcp.exa.ai/mcp, no API key needed to start; open source (exa-mcp-server)\"\n Agent extras: \"Agent Skills (npx skills add exa-labs/agent-skills), Claude connector\"\n brave-search-api:\n Endpoints: \"Web, news, images, videos, local and LLM context; Answers; spellcheck and autosuggest\"\n Index: \"Own independent index, 40+ billion pages\"\n Search price: \"$5 per 1,000 requests; Answers $4 per 1,000 queries plus tokens\"\n Free allowance: $5 of credits every month\n MCP server: \"Yes, brave-search-mcp-server (stdio by default, HTTP optional)\"\n Agent extras: Goggles to re-rank or exclude domains\n parallel-api:\n Endpoints: \"Search, Extract, Task (deep research), Responses, Monitor, FindAll\"\n Index: Not stated in the sources used\n Search price: \"$0.001 to $0.005 per request for 10 results ($1 to $5 per 1,000)\"\n Free allowance: \"Up to 5,000 requests per month free; $5 monthly credit and up to $80 at signup\"\n MCP server: Yes (Search MCP and Task MCP)\n Agent extras: \"Compressed excerpts sized for LLM context; structured outputs with citations and confidence on Task\"\n perplexity-api:\n Endpoints: \"Search API (raw results), Sonar and Agent APIs (grounded answers), Embeddings\"\n Index: Not stated in the sources used\n Search price: \"$5 per 1,000 requests, no token costs\"\n Free allowance: Not stated in the sources used\n MCP server: Yes (Perplexity API MCP server)\n Agent extras: Multi-provider Agent API for web-grounded answers\nverdict: \"For raw results to feed your own agent, Exa, Brave and Parallel are the most transparent on price, and Brave is the only one stating index size. Tavily is the easiest all-in-one (search, extract, crawl and research behind one key and an OAuth MCP server). Choose Perplexity or Parallel's Task API when you want the provider to do the research and return a cited answer.\"\nsources:\n - title: Tavily docs index (llms.txt)\n url: https://docs.tavily.com/llms.txt\n accessed: 2026-10-02\n - title: Tavily pricing\n url: https://www.tavily.com/pricing\n accessed: 2026-10-02\n - title: Exa docs index (llms.txt)\n url: https://exa.ai/docs/llms.txt\n accessed: 2026-10-02\n - title: Exa pricing\n url: https://exa.ai/pricing\n accessed: 2026-10-02\n - title: Exa docs, Exa MCP\n url: https://exa.ai/docs/get-started/exa-mcp\n accessed: 2026-10-02\n - title: Brave Search API\n url: https://brave.com/search/api/\n accessed: 2026-10-02\n - title: Brave Search API pricing\n url: https://api-dashboard.search.brave.com/documentation/pricing\n accessed: 2026-10-02\n - title: Brave Search MCP server README\n url: https://github.com/brave/brave-search-mcp-server\n accessed: 2026-10-02\n - title: Parallel pricing\n url: https://parallel.ai/pricing\n accessed: 2026-10-02\n - title: Parallel docs index (llms.txt)\n url: https://docs.parallel.ai/llms.txt\n accessed: 2026-10-02\n - title: Perplexity docs, Pricing\n url: https://docs.perplexity.ai/docs/getting-started/pricing\n accessed: 2026-10-02\n - title: Perplexity docs index (llms.txt)\n url: https://docs.perplexity.ai/llms.txt\n accessed: 2026-10-02\nrelated: [research-agent, connect-an-agent-to-a-remote-mcp-server, coding-agent-starter]\nlast_verified: 2026-10-02\npublished: 2026-10-02\n---\n\n## What these APIs are for\n\nA model's training data is months old, and a general web search engine returns pages built for people. These APIs return results built for a model: clean text or excerpts, sometimes full page contents, in one call, priced per request. All five offer an MCP server, so a harness such as Claude Code or Codex can use them without custom code; see the [remote MCP guide](https://indexagentica.com/guides/connect-an-agent-to-a-remote-mcp-server/).\n\nThey split into two groups:\n\n- **Search primitives** return ranked results and excerpts, and your agent does the reading and reasoning: [Exa](https://indexagentica.com/entries/exa-api/) Search, [Brave Search](https://indexagentica.com/entries/brave-search-api/), [Parallel](https://indexagentica.com/entries/parallel-api/) Search, the [Perplexity](https://indexagentica.com/entries/perplexity-api/) Search API and [Tavily](https://indexagentica.com/entries/tavily-api/) Search.\n- **Research endpoints** run a multi-step search and return a cited answer or a structured result: Tavily Research, Exa Answer and Exa Agent, Parallel Task, Perplexity Sonar and Agent APIs, and Brave Answers. They save agent turns but hide the intermediate steps, which makes them harder to audit.\n\n## Choosing\n\n- **Price per search** ranges from about $1 to $5 per 1,000 basic requests among the vendors that publish per-request prices; deeper modes, more results and page contents add to that. Tavily prices in credits whose cost depends on the endpoint and depth.\n- **Index.** Brave and Exa run their own indexes, and Brave publishes a size (40+ billion pages, with 100+ million page updates a day). The others don't state where results come from in the pages used here.\n- **Reading pages.** If your agent needs full text, check whether contents come in the same call (Exa Contents, Tavily Extract, Parallel Extract) or need a separate reader such as [Firecrawl](https://indexagentica.com/entries/firecrawl-api/) or [Jina Reader](https://indexagentica.com/entries/jina-reader/).\n- **MCP auth.** Exa's hosted server works without a key to start, and Tavily's supports OAuth, which keeps keys out of config files. Brave's server now defaults to stdio and needs `BRAVE_MCP_TRANSPORT=http` for HTTP.\n- **Evaluate on your own queries.** Quality varies by domain (news, code, academic, local). Run a fixed set of 20 to 50 real queries through two or three APIs and compare what your agent does with the results.\n\nFor a full research setup built on these, see the [research agent stack](https://indexagentica.com/stacks/research-agent/).\n", "html": "

What these APIs are for

\n

A model's training data is months old, and a general web search engine returns pages built for people. These APIs return results built for a model: clean text or excerpts, sometimes full page contents, in one call, priced per request. All five offer an MCP server, so a harness such as Claude Code or Codex can use them without custom code; see the remote MCP guide.

\n

They split into two groups:

\n\n

Choosing

\n\n

For a full research setup built on these, see the research agent stack.

" }