comparison

ScrapeNest as a Tavily alternative

Tavily is a search API for AI agents with extract, crawl, map and research endpoints. It joined Nebius in February 2026. This page sets its pricing, free tier, endpoints, MCP server and SDKs next to ScrapeNest's, so you can judge the fit for your own workload.

Tavily and ScrapeNest at a glance

Topic Tavily (November 2026) ScrapeNest
Billing API credits on monthly plans; credits reset on the first day of each month. Pay-as-you-go at $0.008 per credit once a plan's credits run out. source Credits from one balance for every endpoint. Monthly plans, or pay-as-you-go at $0.0002 per credit. details
Free tier Researcher: 1,000 API credits a month. No credit card required. source 1,000 credits a month. No card required. details
Entry paid plan Project: $30/month for 4,000 API credits. source Starter $29/month for 200,000 credits. details
Other plans Bootstrap $100/month (15,000 credits), Startup $220 (38,000), Growth $500 (100,000), Enterprise custom. source Growth $99/month for 1,500,000 credits; Pro $249/month for 4,000,000 credits; Scale $599/month for 10,000,000 credits. details
Unit costs Basic search 1 credit, advanced search 2. Extract 1 credit per 5 successful URLs (basic). Map 1 credit per 10 pages. Research 4 to 250 credits per request. source Page fetch 1 credit (plain HTTP), 5 (browser), 10 (residential) or 40 (residential with a browser). Web search 5. Cited answer 20 + 5 per source. Maps reviews 100 per 50 reviews. Transcripts 20 per 2 minutes. details
Main endpoints Search, extract, crawl, map, research and usage. source Web search, deep search, cited answers, any-URL scraping, batch scraping (10 URLs), screenshots, PDFs, Google Maps reviews (up to 10,000 per place), gas prices, audio transcripts, translation. details
MCP server Hosted at mcp.tavily.com (OAuth, an Authorization header or a key in the URL), plus the tavily-mcp npm package. source Hosted at scrapenest.dev/mcp with nine tools; API key in a Bearer header, or OAuth from ChatGPT. details
Official SDKs Python (tavily-python) and JavaScript (@tavily/core), plus a CLI. source Python and TypeScript/JavaScript (scrapenest 0.5.2). details

Where Tavily is stronger

  • Crawl and map endpoints for walking a site. ScrapeNest has neither. source
  • A research endpoint with streaming, JSON Schema output and citation formats such as APA, MLA and Chicago. source
  • Keyless access: search and extract work without an API key, free and rate-limited. source
  • OAuth sign-in on its hosted MCP server for clients such as Claude Code. ScrapeNest's OAuth covers ChatGPT only; other clients use an API key. source
  • Integration guides for LangChain, LlamaIndex, CrewAI, n8n, Zapier and the Vercel AI SDK, among others. source

Where ScrapeNest is different

  • A web search is 5 credits and merges several engines into one ranked list: Bing, DuckDuckGo and Brave by default, with Google added at depth=thorough.
  • Deep search returns the results plus the full text of the top pages in one call, and the answer endpoint returns a written answer with numbered citations.
  • Any-URL scraping with browser rendering, Markdown output, CSS extraction rules and AI extraction.
  • Google Maps reviews (up to 10,000 per place), audio transcripts, screenshots, PDFs, translation and US gas prices.
  • Monthly plans from $29 for 200,000 credits, or pay-as-you-go at $0.0002 per credit.

Tavily features and the closest ScrapeNest endpoint

If you are moving a workload, start here. The request and response formats differ, so plan to change your client code.

TavilyScrapeNestNotes
/search POST /v1/search time_range, include_domains, exclude_domains and topic filters.
/search with content POST /v1/search/deep Results plus the readable text of the top 1 to 10 pages.
Answers POST /v1/search/answer Numbered [n] citations with the source list.
/extract POST /v1/scrape/url or /v1/scrape/batch One URL, or up to 10 per batch call.
/research POST /v1/search/answer A single cited answer; no multi-step research runs.
/crawl and /map No equivalent Collect URLs yourself, then fetch them with batch scraping.

Try the same job on ScrapeNest

This sample uses the Python SDK (pip install scrapenest); the TypeScript SDK and plain HTTP work the same way. The closest ScrapeNest calls for search and cited answers:

import os
from scrapenest import Client

client = Client(os.environ["SCRAPENEST_API_KEY"])

hits = client.search("postgres 18 release notes", num_results=10)
for hit in hits.results:
    print(hit.position, hit.title, hit.url, hit.engines)

answer = client.search_answer("What changed in PostgreSQL 18?", max_sources=5)
print(answer.answer)
for citation in answer.citations:
    print(citation.index, citation.url)

questions

ScrapeNest vs Tavily: questions

Is ScrapeNest a search API for AI agents like Tavily?

Yes, in part. ScrapeNest has web search, deep search (results plus page text) and cited answers, and the same calls are available to AI assistants through its MCP server. It has no crawl, map or multi-step research endpoint.

What does a search cost on each?

As of November 2026, a basic Tavily search costs 1 API credit and an advanced search 2; Tavily's Project plan is $30 a month for 4,000 credits. A ScrapeNest web search costs 5 credits, and ScrapeNest's Starter plan is $29 a month for 200,000 credits.

Do both have an MCP server?

Yes. Tavily hosts one at mcp.tavily.com and publishes the tavily-mcp package. ScrapeNest hosts one at scrapenest.dev/mcp with nine tools.

Sources

Every Tavily fact on this page comes from these pages, checked on November 4, 2026:

ScrapeNest facts come from the pricing page, the credit table and the endpoint catalog. Spotted something out of date? Email [email protected].

More comparisons

1,000 free credits a month

Compare it on your own pages

Start with 1,000 free credits a month: sign up, verify your email, then mint a key. No card required.