Every PageCrawl.io plan includes AI credits that work automatically with no setup. For most users, the included credits are all you need. This guide is for users who bring their own API key (BYOK) and want to choose a specific model, from budget options to the most capable models, with cost comparisons.
Using included credits? You don't need to choose a model. PageCrawl automatically uses optimized models on your behalf. Each 4,000-token block costs 1 credit (Basic tier) or 10 credits (Pro tier, Ultimate plan only). See AI-Powered Change Detection for details on how credits work.
Note: The models and prices in this guide were checked against each provider's own model and pricing pages on 6 November 2026. Prices change often, so verify current rates on the OpenAI pricing page, the Gemini API pricing page, and the Anthropic pricing page.
Which AI providers can I use with my own key?
PageCrawl.io supports three providers for Bring Your Own Key: OpenAI, Google Gemini, and Anthropic Claude. OpenAI is pre-selected when you set up a key, but you can pick any of the three. Each provider has a default model, marked with a star in the model picker, that is a sensible starting point for everyday monitoring.
The model you pick summarizes each change, assigns its importance score, and helps separate meaningful updates from noise.
- OpenAI: the GPT-6 family, with GPT-6 Luna as the default
- Google Gemini: Gemini 4 Argon and the Gemini 3.x Flash models, with Gemini 3.8 Flash as the default
- Anthropic Claude: the Claude 5.5 family, with Claude Haiku 5.5 as the default
How much does AI cost per check with my own key?
With your own key, you pay your provider per token. A typical check sends 1,000 to 5,000 input tokens and gets back a summary of about 150 output tokens, so a check on a default model costs a fraction of a cent. AI only runs when a page actually changes, not on every check.
What is a token?
A token is roughly 4 characters or about 3/4 of a word. AI providers charge based on tokens processed:
- "Hello world" = ~3 tokens
- A typical paragraph = ~100 tokens
- A blog post (1,000 words) = ~1,300 tokens
- A full webpage = ~2,000-10,000 tokens
How PageCrawl uses tokens
Most of the tokens in a check are input tokens (the page content sent to AI). Output is short, because summaries are typically just 1-2 paragraphs (~100-200 tokens).
Typical token usage per check:
- Simple page (blog post, article): ~1,000-2,000 tokens
- Medium page (product page, news): ~2,000-5,000 tokens
- Large page (documentation, e-commerce): ~5,000-10,000 tokens
Example costs per check
These examples assume 150 output tokens per check and use the prices listed in this guide.
| Model | Price per 1M tokens (input / output) | 2,000-token page | 5,000-token page |
|---|---|---|---|
| GPT-6 Luna | $0.10 / $0.50 | ||
| Gemini 3.8 Flash (through 31 December 2026) | $0.75 / $3.75 | ||
| Gemini 3.8 Flash (from 1 January 2027) | $1.50 / $7.50 | ||
| Claude Opus 5.5 | $4 / $20 |
Output tokens cost about five times as much as input tokens on these models, so on a small page the short summary can add a third to the cost of a check. Models that reason before answering also bill those reasoning tokens as output, so real costs can be higher than these estimates. AI only runs when a meaningful change is detected on the page, and tiny, insignificant changes are filtered out before they reach AI, so you only spend tokens on changes that matter.
Which models can I choose from?
The model picker groups each provider's models into Latest models and Older models. The latest group holds the current lineup listed below, and the star marks each provider's default. Older models stay selectable for workspaces that already use them. Prices are in US dollars per million input and output tokens.
OpenAI models
| Model | Price (input / output) | Notes |
|---|---|---|
| GPT-6 Astra | $10 / $50 | Most capable OpenAI model, highest price |
| GPT-6.1 Sol | $2 / $10 | Near-Astra performance at a lower cost, according to OpenAI |
| GPT-6.1 Terra | $1 / $5 | Mid-priced option |
| GPT-6 Luna ⭐ | $0.10 / $0.50 | Default, lowest price in the current lineup |
Older models: GPT-6 Sol, GPT-5.6 Sol, Terra and Luna, GPT-5.5, GPT-5.4, GPT-5.4 Mini and Nano, the earlier GPT-5.x models, GPT-4.1 and GPT-4o.
Google Gemini models
| Model | Price (input / output) | Notes |
|---|---|---|
| Gemini 4 Argon | $2 / $10 through 30 November 2026, then $4 / $20 | Google's flagship reasoning model, most capable |
| Gemini 3.8 Flash ⭐ | $0.75 / $3.75 through 31 December 2026, then $1.50 / $7.50 | Default, Google's most intelligent Flash model |
| Gemini 3.5 Flash-Lite | $0.30 / $2.50 | Low-cost option |
| Gemini 3.1 Flash-Lite | $0.25 / $1.50 | Budget, lowest price in the current lineup |
Older models: Gemini 3.1 Pro, Gemini 3 Flash and Gemini 3 Pro.
Gemini 2.5 models are no longer offered. Google restricts them to existing users, and from 1 December 2026 Google routes gemini-2.5-flash requests to Gemini 3.5 Flash-Lite. If your workspace still uses a Gemini 2.5 model, pick a current one.
Anthropic Claude models
| Model | Price (input / output) | Notes |
|---|---|---|
| Claude Opus 5.5 | $4 / $20 | Most capable Anthropic model, highest price |
| Claude Sonnet 5.5 | $2 / $10 | Mid-priced option |
| Claude Haiku 5.5 ⭐ | $0.10 / $0.50 for prompts up to 100K tokens ($0.50 / $2.50 above) | Default, built for high-volume, latency-sensitive tasks |
Older models: Claude Sonnet 5, Claude Opus 5, the Claude Opus 4.x and Sonnet 4.x models, and Claude Haiku 4.5.
Which model should I pick for my pages?
Start with your provider's default (GPT-6 Luna, Gemini 3.8 Flash, or Claude Haiku 5.5) and move up a tier only for pages where the summaries miss detail. PageCrawl.io only calls AI when a page actually changes, so if you monitor 1,000 pages and 150 change, you pay for 150 AI requests, not 1,000.
The AI settings page offers three Quick Select cards that set the model in one click:
| Quick Select card | OpenAI | Anthropic | Google Gemini |
|---|---|---|---|
| 🚀 Best / Most Capable | GPT-6.1 Sol | Claude Opus 5.5 | Gemini 4 Argon |
| ⭐ Good Quality | GPT-6 Luna | Claude Haiku 5.5 | Gemini 3.8 Flash |
| 💰 Budget | GPT-6 Luna | Claude Haiku 5.5 | Gemini 3.1 Flash-Lite |
- 🚀 Best / Most Capable suits complex pages where accuracy matters most, such as legal documents, terms of service, and compliance monitoring. These models cost significantly more, so use them only where precision is critical.
- ⭐ Good Quality is the starting point for most BYOK users and handles the vast majority of everyday monitoring.
- 💰 Budget keeps costs lowest for high-volume monitoring of simple pages such as blog posts, news, and documentation. For OpenAI and Anthropic it picks the same model as Good Quality, because GPT-6 Luna and Claude Haiku 5.5 are already the lowest-priced current models from those providers.
Starting points by content type
| Content type | Start with | Move up to if summaries miss detail |
|---|---|---|
| Blogs, news, docs | GPT-6 Luna, Claude Haiku 5.5, or Gemini 3.1 Flash-Lite | GPT-6.1 Terra or Gemini 3.8 Flash |
| E-commerce, pricing | GPT-6 Luna or Gemini 3.1 Flash-Lite | Gemini 3.8 Flash |
| Legal, ToS, compliance | Claude Haiku 5.5 or Gemini 3.8 Flash | Claude Sonnet 5.5, Claude Opus 5.5, GPT-6.1 Sol, or Gemini 4 Argon |
| Competitor monitoring | GPT-6 Luna or Claude Haiku 5.5 | GPT-6.1 Terra or Claude Sonnet 5.5 |
| API docs, changelogs | GPT-6 Luna or Gemini 3.8 Flash | GPT-6.1 Sol |
What will BYOK cost me in a month?
Monthly cost depends on how many of your pages change, how large they are, and the model you pick. Multiply the number of changes you expect by the cost per check of your model. Start on a low-cost default model and set a monthly request limit so a busy month cannot produce an unexpected bill.
Costs can vary significantly. These are estimates only. Your actual costs depend on:
- Page complexity and content length
- How often pages change
- Deep Analysis setting (on = full page, off = changes only)
- Max token settings
Token usage by page type:
- Simple pages (blogs, docs): ~500 tokens
- Average pages: ~2,000 tokens
- Content-heavy pages: ~5,000-10,000 tokens
- Complex pages (e-commerce, SPAs): 10,000-25,000+ tokens
Recommendation: Start with a low-cost model like GPT-6 Luna, Claude Haiku 5.5, or Gemini 3.1 Flash-Lite and set strict monthly limits to avoid unexpected bills.
Controlling token usage
You can reduce token usage in PageCrawl.io settings:
- Deep Analysis off: Only send changed text to AI (lower tokens, less context)
- Deep Analysis on: Send entire page for better understanding (higher tokens)
- Max tokens limit: Set a maximum per request (falls back to diff if exceeded)
- Monthly request limits: Set max AI requests per month to cap costs
- Per-page daily limit: Prevent noisy pages from consuming all your AI budget
Note: Check your actual token usage in PageCrawl.io's AI statistics to estimate your costs accurately.
How do I set up BYOK in PageCrawl.io?
Get an API key from OpenAI, Google Gemini, or Anthropic, then open Settings > Workspace > Integrations, open the AI Features card, and add the key in the Bring Your Own Key section. OpenAI is pre-selected; choose your provider and model, test the key, and save. AI credits stop being consumed once your key is saved.
Step 1: Get your API key
| Provider | Get Key At |
|---|---|
| OpenAI | platform.openai.com > API Keys |
| Google Gemini | ai.google.dev > Get API Key |
| Anthropic | console.anthropic.com > API Keys |
Step 2: Configure in PageCrawl.io
- Go to Settings > Workspace > Integrations
- Click Manage or Setup on the AI Features card
- Scroll down to the Bring Your Own Key section
- Select your AI provider (OpenAI is pre-selected)
- Choose a model from the dropdown (the star marks the provider's default)
- Click Add API Key, paste your key, and use Test Key to verify it works
- Click Save Configuration
Step 3: Use Quick Select for easy model choices
After adding your API key, the settings page shows three Quick Select cards:
- 🚀 Best / Most Capable - Most accurate results, higher cost
- ⭐ Good Quality - The starting point for most users
- 💰 Budget - Lowest cost, good for simple monitoring
Click a card to switch to that model. You can also choose any model from the dropdown, including the Older models group.
Step 4: Optimize with model overrides
You can customize AI models at three levels:
- Workspace default - applies to all pages
- Template override - applies to pages using that template
- Page override - applies to individual pages
Overrides choose from the models of the provider your workspace key belongs to.
Example strategy with an Anthropic key:
- Workspace default: Claude Haiku 5.5 (lowest cost)
- Legal template: Claude Sonnet 5.5 (longer, denser text)
- Critical page: Claude Opus 5.5 (most capable)
With an OpenAI key the same ladder is GPT-6 Luna, GPT-6.1 Terra or Sol, and GPT-6 Astra. With a Gemini key it is Gemini 3.1 Flash-Lite or Gemini 3.8 Flash, then Gemini 4 Argon.
How can I keep BYOK costs down?
Start on your provider's default or Budget model, set a monthly request limit, and keep Deep Analysis off unless a page needs full context. Check frequency does not change AI cost, because AI only runs when a page changes. Use templates to give a more capable model only to the pages that need it.
- Start with the Good Quality tier - GPT-6 Luna, Gemini 3.8 Flash, or Claude Haiku 5.5 are each provider's default
- Use templates - Group similar pages with the same model to optimize costs by content type
- Check frequency doesn't affect AI costs - AI only runs when changes occur, not on every check
- Set monthly limits - Use the "AI Requests Per Month" setting to cap spending
- Monitor token usage - Check the token statistics in your AI settings to understand your actual costs
- Use per-page daily limits - Prevent frequently updating pages from consuming all your budget
- Watch for dated price changes - Gemini 4 Argon's price doubles after 30 November 2026 and Gemini 3.8 Flash's after 31 December 2026
How do the providers handle my data?
With BYOK, PageCrawl sends the content needed for each AI feature directly to the provider you chose, under that provider's API terms. OpenAI and Anthropic state that API data is not used for training by default. Google's terms differ between the free and paid Gemini API tiers, so review them before monitoring sensitive content.
| Provider | Data Usage | Best For |
|---|---|---|
| OpenAI / Anthropic | API data not used for training by default | Confidential content, legal docs |
| Google Gemini | Review Google's Gemini API terms for your tier | General monitoring |
When using included AI credits, content is processed through PageCrawl's managed AI infrastructure. When using BYOK, content is sent directly to your chosen provider.
Data protection policies: OpenAI, Anthropic, Google
Privacy note: Free API tiers can have different data terms from paid tiers. Use a paid tier for sensitive content.
FAQ
Do I need BYOK to use AI? No. All plans include AI credits that work automatically. BYOK is optional for users who want unlimited usage or specific model control.
Which providers does BYOK support? OpenAI, Google Gemini, and Anthropic. OpenAI is pre-selected when you set up a key.
Which model is selected by default? Each provider's default is marked with a star in the picker: GPT-6 Luna for OpenAI, Gemini 3.8 Flash for Google Gemini, and Claude Haiku 5.5 for Anthropic.
What happens when my credits run out? Page monitoring continues normally, but AI summaries pause until credits reset next month. You can also switch to BYOK for unlimited usage.
Can I switch between credits and BYOK? Yes, at any time in Settings > Workspace > Integrations > AI.
Can I switch models after starting? Yes. The new model is used from the next check that finds a change. Historical data remains intact.
Do I pay for checks that don't find changes? No. AI only runs when pages actually change.
Can I use different models for different pages? Yes, via workspace defaults, template overrides, and page-level overrides, choosing from the models of your workspace's provider.
Why are Gemini 2.5 models missing from the picker? Google restricts Gemini 2.5 models to existing users, and from 1 December 2026 Google routes gemini-2.5-flash requests to Gemini 3.5 Flash-Lite. Gemini 3.8 Flash and the Gemini 3.x Flash-Lite models replace them.
Related Articles
- AI-Powered Change Detection and Smart Filtering - Learn how AI summarization and Importance Scoring work
- AI Integration Setup Guide (BYOK) - Step-by-step guide to configure your API keys