Subscription usage by project
The dashboard now reports subscription usage by project, so you can see how it is spread across your projects.
Changelog · 2025-12
The launch of QuantSearch AI, static site deployment from templates, AI search crawling across the platform, subscription usage by project, a new Python API client, and SDK 4.8 releases.
8 updates across 7 products shipped this month.
The dashboard now reports subscription usage by project, so you can see how it is spread across your projects.
A new Deploy site option on the dashboard home deploys a static site from one of Quant's static site generator templates. It uses your organisation's GitHub App connection. The dashboard creates the project or uses an existing one, and creates a repository from the template in your GitHub account. It then adds the deployment variables and secrets, and tracks the first GitHub Actions build.
Related documentation: Start from a template: Create a static site from a template
When origin protection is on for a container, you can now list IP addresses or CIDR ranges, such as your VPN exit addresses, that can reach the application directly. Set the list in the container settings or with the originProtectionConfig.ipAllow field in the API. Requests that come through Quant are allowed as before.
QuantSearch AI (quantsearch.ai) is a new hosted service that adds AI search to any website. It has its own sign-up and accounts, separate from the Quant dashboard and from AI Search in Quant projects. You add your site, and QuantSearch AI crawls and indexes its pages. Visitors search by meaning instead of exact keywords and can get direct AI answers to their questions. You add a chat widget, a modal search or a search page to your site with a short script. Search groups combine several sites into one search, and an analytics page shows what visitors search for. API keys give access to the search and ingestion API, which the documentation on quantsearch.ai describes.
A new Python API client is on PyPI, generated from the same unified API specification as the TypeScript, PHP and Go clients.
AI search support is now available across the platform. You can turn on AI Search for a project from its Search page in the dashboard, as an alternative to keyword search. When browser mode is on, the crawler executes JavaScript automatically for AI search crawls. For projects without AI Search, a new Enable Progress Tracking crawler setting tracks each crawl's pages discovered, pages crawled and errors. You can follow crawls in a new crawl status tracker and preview your AI search results in the dashboard. You can also add content to the AI search index directly through the API, without a crawl.
Related documentation: Get started with AI Search: Enable AI Search
The TypeScript, PHP, Python and Go clients were updated to v4.6.0 and then v4.8.0, generated from the unified API specification. In the TypeScript, PHP and Go clients, version 4.6.0 replaces the single AI services class with separate classes for AI agents, custom tools, inference, models, monitoring, sessions, tools and the vector database. Version 4.6.0 also adds DNS validation and go-live records to domains. Version 4.8.0 adds the origin protection IP allow list for containers.
Change code that uses the single AI services class to use the new AI classes when you upgrade to 4.6.0 or later.
Some crawls did not finish, or showed wrong queue counts. Crawls now finish and show the correct counts. Network errors now show a clear status. Pages that return an error during a crawl are now tracked and counted as errors in the crawl status tracker. Exclude rules now match against only the path and the query string, not the full URL. A pattern such as ^/admin/.* now excludes those pages, and a pattern that includes the domain no longer matches.
Remove the domain from crawler exclude patterns, because they now match only the path and query string.
No updates for that product this month.