Zero API calls. Zero server costs. Instant search.
Sprigr compiles your search index into a blazingly fast client-side search engine. Users search locally in the browser. No per-query fees, no infrastructure to manage, no servers to scale. Your search cost is the same at 10 searches per month or 10 million.
Sprigr compiles your search index into a client-side engine. Search runs locally in the browser with zero API calls. Your cost is a flat monthly fee regardless of query volume. Whether your users run 10 searches or 10 million, the bill doesn't change.
You focus on the work. Sprigr runs the paperwork.
- Zero per-query costs
- No infrastructure
- Sub-10ms on typical indexes
- Works offline
Performance without compromise
Eliminating API calls does not just save money. It makes search fundamentally faster and more reliable.
-
No cold starts
Serverless search functions take 50-500 ms to cold start. Sprigr's search engine is already loaded in the browser. The first search query is just as fast as the thousandth, under 10 ms on typical indexes. There is no waiting for a function to spin up, no Lambda initialization, no container warm-up.
-
No rate limits
API-based search enforces rate limits to protect server resources. When your site gets a traffic spike (a product launch, a viral post, a Black Friday rush) rate limits throttle your search right when it matters most. Sprigr has no rate limits because there is no server to protect. Each user runs search independently in their browser.
-
Works offline
Once the search engine and index are cached, search works without any network connection. Users on planes, trains, or spotty mobile connections can still search your content. This is critical for progressive web apps, mobile apps, and documentation sites used in areas with unreliable internet.
-
Flat cost at any scale
Per-query pricing punishes growth. Sprigr costs the same whether your users run 10 searches or 10 million. Your search bill is predictable and doesn't surge with traffic.
-
Global edge delivery
The compiled index is served from a 300+ location edge network. First-load time is optimized regardless of where your users are. After that, the engine is cached locally.
-
Privacy preserved
Keyword queries are answered on the device and are not sent to Sprigr. Analytics are optional and off unless you post events to /1/events; if you do, those events are stored server-side. Your records are hosted on Cloudflare, isolated per tenant, and Sprigr never trains on them.
How zero-API search works
Three steps from your data to instant, serverless search.
- 01
Push your data
Send records to the Sprigr REST API. Products, articles, documentation, job listings, any structured JSON. Records are stored on Cloudflare, isolated per tenant.
- 02
Compile to a client-side engine
Sprigr builds a compact binary search index optimized for in-browser execution. Indexes with semantic search enabled are larger. The index is served from a global edge network.
- 03
Search locally
The lightweight search engine loads in the user's browser and searches the index locally. Keyword queries return in under 10 ms on typical indexes. No API call per search, no cost per query. Hybrid semantic mode fetches one embedding per query from /1/embed.
At 100,000 searches/mo, Algolia is around $50. At 1 million, around $500. At 10 million, around $5,000. Sprigr is flat: whether your users run 10 searches or 10 million, the bill doesn't change. Predictable pricing that rewards growth instead of punishing it.
See pricing →Questions
How does Sprigr search work without API calls?
You push your data to Sprigr via a REST API during indexing. Sprigr compiles that data into a compact compiled binary search index. When a user visits your site, the search engine and index are loaded into the browser. Keyword queries run locally in the browser with no API call per search. The network requests are the initial load of the engine and index (cacheable with normal HTTP headers), one embedding call per query if you enable hybrid semantic mode, and analytics events only if you choose to send them.
Is there a limit to how many searches users can run?
No. Since searches run in the user's browser, there is no server to rate-limit. Users can search as fast as they can type, and every keystroke produces results in under 10 milliseconds on typical indexes. There are no per-query costs, no daily quotas, and no throttling. The only limit is the user's device performance, which on modern hardware is effectively unlimited for search workloads.
What happens if my site gets a traffic spike?
Nothing changes for search performance. Each user's searches run in their own browser, so search scales linearly with users at zero marginal cost. The search engine is served from a global edge network across 300+ locations, so the initial load is fast regardless of traffic volume. No auto-scaling, no capacity planning, no emergency infrastructure bills.
How large can my search index be?
The practical limit depends on record size and the device: the whole index has to fit comfortably in browser memory. Indexes with semantic search enabled include vector embeddings and are larger. Large indexes are automatically split into shards loaded in parallel. For very large datasets, Sprigr also provides a server-side REST API as a fallback.
Does Sprigr work offline?
Keyword search does. Once the engine and index are loaded, keyword results, filtering, and highlighting all work without a network connection, which suits progressive web apps, mobile apps, and documentation sites in places with unreliable connectivity. Hybrid semantic mode needs a network call per query for the embedding.
Instant search with zero API calls.
Paid plans from $49 a month.