> **For coding agents and LLMs:** This is one Social Fetch marketing page (markdown export). For API orientation start with [`/llms.txt`](https://www.socialfetch.dev/llms.txt); for auth and crawl rules use [`/agents.txt`](https://www.socialfetch.dev/agents.txt); for every operation use [`/llms-endpoints.txt`](https://www.socialfetch.dev/llms-endpoints.txt). ## This page - **On-site (HTML):** - **Markdown (.mdx) URL:** --- # Social Fetch vs Exa Exa searches the web with semantic and neural retrieval — people, companies, code, and general queries with contents included on search. Social Fetch looks up social network entities you already identified. Both show up in AI stacks; only one returns a TikTok followerCount field. ## At a glance | | Social Fetch | Exa | | --- | --- | --- | | Primary job | Structured social media lookups | Neural / semantic web search for AI | | Pricing (headline) | From $1.35 per 1,000 social lookups | $7 per 1,000 searches (up to 10 results + contents) | | Free tier | 100 credits, never expire | Generous free credits / monthly allowance (see Exa pricing) | | People / company search on the web | Not the product | Core strength | | TikTok profile JSON | Dedicated route + normalized schema | Search results — not a social API | | Social catalog | 264 ops / 25 platforms | Web index — category mismatch | ## Pricing (last verified 2026-07-21) **Search requests vs social operations.** Exa: Search $7/1k (≤10 results with text), Deep Search $12, Contents $1/1k pages. Social Fetch Scale: $1.64/1k social lookups. A thousand Exa searches ≠ a thousand Instagram profiles. | | Social Fetch | Exa | Note | | --- | --- | --- | --- | | 1,000 search requests | N/A | $7 (standard Search, ≤10 results) | | | 1,000 social profile lookups | ~$1.64 | N/A as structured social | | | Full page contents add-on | Web fetch is a separate social-adjacent utility | $1 per 1,000 pages on Contents endpoint | | | Billing style | Prepaid packs; PAYG never expires | Usage-based API credits | | Exa pricing from exa.ai/pricing (July 2026). Free credit promotions change; verify the current signup and monthly free allotment on Exa's site before modeling a prototype budget. ## Feature matrix | Feature | Social Fetch | Exa | | --- | --- | --- | | Synchronous REST JSON (no actor run polling) | Yes | Yes | | One API key for all platforms | Yes | Yes | | Unified schema across platforms | Yes | No | | Public OpenAPI specification | Yes | Partial | | Official TypeScript SDK | Yes | Yes | | Interactive API playground | Yes | Partial | | TikTok, Instagram, YouTube, LinkedIn, Facebook, Reddit, X in one vendor | Yes | No | | Pay-as-you-go without required monthly subscription | Yes | Yes | | PAYG credits never expire | Yes | Partial | | Web fetch (markdown/HTML) in same product | Yes | Partial | | Scheduled account / feed monitors | Yes | Partial | | Signed webhooks when watched accounts post | Yes | No | ## Exa finds; Social Fetch fetches known social entities Exa shines when the user (or agent) has a fuzzy need: find similar companies, research a person across the web, pull code or paper context. Contents and highlights are tuned for RAG token budgets. Deep Search trades latency and dollars for harder queries. Social Fetch shines when the identifier is already known: a handle, a post URL, a channel ID. The hard part is not retrieval ranking — it is keeping platform extraction working and returning a stable schema. Pointing Exa at "instagram.com/foo" and parsing the snippet is a prototype trick, not a production social integration. ## Wire the right tool to the right argument type In an MCP or tool-calling setup, give the model Exa (or Tavily) when the argument is a natural-language query. Give it Social Fetch when the argument is a social URL or handle. Mixing them — Exa for everything — recreates the HTML-parsing tax Firecrawl users hit on social pages. Exa's people and company search can still help discovery before you have a handle. Once you do, hand off to Social Fetch for live metrics. ## Seven dollars per thousand is fine — for search Exa at $7/1k is priced for search quality, not for competing with $1.64 social lookups. Teams that try to "save money" by using only Exa for creator dashboards usually spend more on LLM tokens cleaning messy snippets than they would on a social API — and still lack typed fields for sorting and filtering. Conversely, do not buy Social Fetch hoping it replaces Exa's neural search. Different index, different product. ## When Exa is the better choice - Semantic web search, people/company discovery, code or research queries - RAG pipelines that need ranked contents and highlights - Agent frameworks already integrated with Exa's SDKs ## When Social Fetch is the better choice - Live social profiles, posts, comments, transcripts as JSON - Dashboards and CRM enrichment keyed by social handles - One envelope across many social networks ## Verdict: Neural search vs social entity API Exa is for finding and reading the web. Social Fetch is for structured social entities you already named. Keep Exa in the research tool slot. Keep Social Fetch in the social data tool slot. Ranking them on a single price-per-1k chart without separating those jobs misleads procurement. ## FAQ ### Is Exa cheaper than Social Fetch? For web search, Exa's $7/1k is the relevant number. For social profile lookups, Social Fetch's $1.64/1k is the relevant number. They are not interchangeable SKUs. ### Can Exa return TikTok follower counts reliably? Only insofar as some web page mentions them. That is not the same as a live profile lookup with typed fields and lookupStatus. ### Should my agent include both tools? Usually yes if it both researches the web and enriches social accounts. Route by argument type: query → Exa; handle/URL → Social Fetch.