Summary
| API | Price | Use case | Reasoning | Type | Latency |
|---|
| Search | $ | Page and excerpt retrieval | - | Synchronous | 1-3s |
| Extract | $ | Page content retrieval | - | Synchronous | 1-20s |
| Chat | $ | Grounded chat completions | Low | Synchronous | 1-3s |
| Task | $-$$$$ | Deep research, enrichment, custom research | Low-High | Asynchronous | 10s - 2hr |
| FindAll | $-$$$$ | List / database building | Low-High | Asynchronous | 10s - 2hr |
| Monitor | $ | Always-on web monitoring | Low | Asynchronous | Ambient |
Search API
By default, the Search API returns 10 page results and their excerpts per request.
| Component | Cost ($/1000) |
|---|
| Per 1,000 requests (default 10 results) | 5 |
| Per 1,000 additional page results & excerpts | 1 |
| Component | Cost ($/1000) |
|---|
| Per 1,000 URLs | 1 |
Web Agents
Chat API
Chat API pricing is based on the model you select. Research models (lite, base, core) are Chat API wrappers over Task API processors and share the same pricing.
| Model | Type | Processor | Cost ($/1000) |
|---|
speed | Simple completions | - | 5 |
lite | Research | lite | 5 |
base | Research | base | 10 |
core | Research | core | 25 |
Task API
Task API pricing is based on the processor you select. Cost is per 1,000 Task Runs. Fast processors have the same pricing as their standard counterparts.
| Processor | Cost ($/1000) | Latency | Strengths |
|---|
lite | 5 | 10s - 60s | Basic metadata, fallback, low latency |
base | 10 | 15s - 100s | Reliable standard enrichments |
core | 25 | 60s - 5min | Cross-referenced, moderately complex outputs |
core2x | 50 | 60s - 10min | High complexity cross referenced outputs |
pro | 100 | 2min - 10min | Exploratory web research |
ultra | 300 | 5min - 25min | Advanced multi-source deep research |
ultra2x | 600 | 5min - 50min | Difficult deep research |
ultra4x | 1200 | 5min - 90min | Very difficult deep research |
ultra8x | 2400 | 5min - 2hr | The most difficult deep research |
| Processor | Cost ($/1000) | Latency | Strengths |
|---|
lite-fast | 5 | 10s - 20s | Basic metadata, fallback, lowest latency |
base-fast | 10 | 15s - 50s | Reliable standard enrichments |
core-fast | 25 | 15s - 100s | Cross-referenced, moderately complex outputs |
core2x-fast | 50 | 15s - 3min | High complexity cross referenced outputs |
pro-fast | 100 | 30s - 5min | Exploratory web research |
ultra-fast | 300 | 1min - 10min | Advanced multi-source deep research |
ultra2x-fast | 600 | 1min - 20min | Difficult deep research |
ultra4x-fast | 1200 | 1min - 40min | Very difficult deep research |
ultra8x-fast | 2400 | 1min - 1hr | The most difficult deep research |
Pricing is per Task Run (row), not per output field (cell). A single Task Run can populate many output fields—whether you request 1 field or 20 fields, the cost is the same.
FindAll API
FindAll API pricing is based on the generator you select, with a fixed cost plus a per-match cost.
| Generator | Fixed Cost | Per Match | Best For |
|---|
preview | $0.10 | $0.00 | Testing queries (~10 candidates) |
base | $0.25 | $0.03 | Broad, common queries where you expect many matches |
core | $2.00 | $0.15 | Specific queries with moderate expected matches |
pro | $10.00 | $1.00 | Highly specific queries with rare or hard-to-find matches |
Monitor API
Monitor requests are priced on a per thousand (CPM) basis for each execution. Execution frequency is defined by the frequency parameter of your monitors (e.g., "1h", "1d", "1w").
| Component | Cost ($/1000) |
|---|
| Per 1,000 requests per frequency | 3 |
