One API call. Finished research. Flat rate.
Never again guess your bill.
FlatSearch is the search infrastructure behind VCP Scanner, a financial equity research platform. You're accessing production capacity at wholesale rates.
Building web research into your agent is a full-time job.
Your agent hits a search API, gets 10 blue links, scrapes each one, strips the HTML, stuffs it into a prompt, runs the LLM, parses the output. You maintain the scraping logic, handle rate limits, deal with blocked pages, and pay for every token of that chain.
At scale, that's not an integration. It's an entire microservice you didn't budget for.
We do the research. You get the answer.
FlatSearch is a managed research sub-agent. Send a question, and we scrape live sources, strip the noise, cross-reference findings, and return a structured JSON answer with cited sources. One API call replaces your scraping loop, content cleaning, and synthesis prompt.
No scraping logic. No token math. Just $1.00 per 1,000 queries.
Two endpoints. One decision: speed or depth.
Your chatbot needs a quick answer, or your agent needs a vetted report. Pick the one that fits.
Fast Search
/v1/search/fastYour chatbot asks a question, we return a cited answer in ~1 second. Built for customer support bots, autocomplete layers, and any agent that needs web context without building a scraping pipeline.
- ~1 second response time
- 3 parallel source scrapes per query
- High-speed synthesis engine
Deep Search
/v1/search/deepHand your agent a complex question — we research it across multiple sources, cross-reference the findings, and return a vetted report. Built for agents where being wrong costs more than being slow.
- 15–25 second reasoning loop
- 5 sources + targeted follow-up scrapes
- Cross-referenced, validated output
Start free. Scale when you're ready.
No lock-in on PAYG. Overages on Pro are billed at standard flat rates — never blocked.
Prototyping and side projects.
- 100 free queries/mo (Fast Search)
- Shared proxy pool access
- Rate limit: 2 req/s
- Discord support channel
Pay only for what you use. Credits never expire.
- Billed at standard flat rates
- Credits persist indefinitely
- Rate limit: 25 req/s
- Standard residential proxy pools
Teams in production. Includes $40 in credits.
- Includes $40 in built-in credits
- Equiv. to 40k Fast or 16k Deep queries
- Overages billed at standard flat rates
- Prioritized proxy routing subnets
- High rate limits (50 req/s)
Dedicated nodes, SLAs, 24/7 support.
- Unlimited custom query capacity
- Dedicated Search Engine nodes
- Custom IP routing & blocklists
- 99.9% service uptime guarantee
- 24/7 dedicated support pager
Developers running 61,000 queries/month save ~$671.
At that volume, traditional variable token APIs cost an estimated $732/month. FlatSearch costs $61. That's not a rounding error — that's a different product category.
Tiers Comparison Matrix
| Feature Specifications | Tier 1: Fast Search | Tier 2: Deep Search |
|---|---|---|
| Retail Pricing | $1.00 per 1,000 queries | $2.50 per 1,000 queries |
| Synthesis LLM Engine | High-speed synthesis engine | Advanced reasoning engine |
| Execution Loop Pipeline | Single-pass context summarization | Multi-stage retry reasoning loop |
| Query Formulation | Generates 1 to 3 concurrent engine checks | Generates 5 initial + targeted secondary refetches |
| Proxy Routing Network | Rotating household residential subnets | Prioritized high-reputation subnets |
| Ingress Content Stripping | Boilerplates, navigation nodes, URL Blocklist | Deep index cross-checking & content gate |
| Standard Rate Limits | 10 requests per second | 25 requests per second |
Every other search API charges you for the tokens we eat.
We got burned by Tavily and Perplexity's token multipliers running our own platform. Variable pricing at our query volume was destroying margin. We built flat pricing because we needed it ourselves.
| API Provider | Core Billing Strategy | True Cost Per 1k Queries |
|---|---|---|
| FlatSearch (Fast Search) | Flat rate wholesale pricing. Zero input/output token tracking. | $1.00 (Flat) |
| FlatSearch (Deep Search) | Flat rate agentic research. Self-critique iterations included. | $2.50 (Flat) |
| Perplexity Sonar | Flat query fee ($6-$14/1k) + variable token usage (input/output/citations). | ~$13.50+ (Variable) |
| Tavily API | Advanced extractions consume up to 250 credits per query. | ~$6.40 (Starter Plan) |
| Exa AI | Credits charged per search link + supplementary fee for summaries. | $7.00 - $15.00 |
| Firecrawl | Capped monthly limits. Targeted AI scrapes cost 5x credits. | $3.20 - $8.30 |
Frequently Asked Questions
Why not just use Google Custom Search + my own LLM?
You can. But then you're maintaining a scraping pipeline, handling rate limits, dealing with blocked pages, cleaning HTML, managing context windows, and paying per-token for the LLM synthesis. We bundle all of that into one API call at a flat rate. Most developers find the engineering time alone isn't worth it.
Why residential proxies? Why does that matter?
Search engines immediately block cloud server IPs with CAPTCHAs. Residential IPs look like real users browsing from home — which they are. This is what lets us achieve 99.9% successful ingestion on queries instead of hitting walls.
What happens if I go over my plan's credits?
We don't cut you off. Overages continue at the standard flat rate and come out of your account balance. No service interruption, no penalty pricing.
Do I only get a summary, or do I get the raw links too?
You get both. Every response includes the synthesized markdown answer AND the raw sources array containing the top 10-25 URLs, titles, and snippets. Use the finished answer for speed, or parse the raw sources if your agent needs total control.