| Primary product type | Developer-first structured web data API (60+ platforms) | Secure browser infrastructure for computer-use agents |
|---|
| Output | Normalized JSON per documented endpoint | A live, controllable browser session — you extract what you need yourself |
|---|
| Agent execution model | Not applicable — fixed endpoints, no agent loop | AI plans the task, then executes as deterministic code — reverts to AI only when runtime requires it |
|---|
| Authentication | API-key auth for Crawlora itself; no third-party site login handling | OmniConnect: end-to-end authentication lifecycle for logging agents into third-party sites |
|---|
| Network layer | Managed proxy routing behind supported endpoints | Built-in Anchor VPN — no third-party proxy dependency |
|---|
| Browser identity | Not applicable | Anchor Chromium: a stealth fork recognized as human, Web-Bot-Auth Verified |
|---|
| Workflow caching | Not applicable | Web Action Cache: maintains workflows as deterministic code after the first AI-planned run |
|---|
| Deployment | Hosted API only | Hosted, or bring-your-own-cloud with your own inference provider |
|---|
| Pricing model | One credit pool shared across all endpoints (Growth: $29/mo for 100,000 credits) | Free: 5 credits/mo (0.1 credit per completed task); Starter: $50/mo for 50 credits, $1.00/credit overage |
|---|
| Agent-native workflows | Hosted MCP tools across the full catalog | SDK, MCP, or any automation platform; agent authentication is the core product |
|---|