Questions about Giant Context
What it is, how you stay in control, what it measures and how to get access.
The basics
The disconnected stack. Building pages by hand, writing and sending emails yourself, a form tool, a rule-based support bot and the connectors between them become apps on one platform that the AI operates.
Only what you switched on. Each step, from ideation through plan and draft to publishing, is a separate switch per content type. Leave publishing off and every draft waits in the queue for you.
On the security page, which covers data residency, AI providers, access, deletion and payments.
How the work gets done
It reads the documents, product notes, earlier marketing pages, screenshots and knowledge base articles in your project, and it writes with the names your documents use.
Pages, blog posts, emails, knowledge base articles and social posts, each grounded in your project's context.
Yes. Ask for a page for a launch, an ad or an email and it is drafted from the same context, with the form and the call to action on it.
Yes. It answers from your knowledge base and website content, and it hands the conversation to a person under the rules you set.
The AI drafts articles for the questions that keep coming up in chat. You approve them.
Yes. Connect Claude, Cursor or any MCP client and operate the platform in plain language, with 215 tools across the eight apps.
What it measures
Opens, clicks and unsubscribes per send. Submissions per form. The questions asked in chat. Your own analytics tag goes in through Google Tag Manager if you want it as well.
It reads what was opened, clicked, submitted and asked, and turns that into ideas in the queue, a page to revise, an article to write, an email for a contact. You approve each one.
Its context does. Every approved page, post and article and every new document in the project is there the next time it writes.
Alpha status
Alpha. Giant Context is a research project exploring automations and agentic workflows. It may become publicly available based on interest.
Expect rough edges and rapid change. Features evolve quickly during alpha.
Not during alpha. A standard SLA is planned for general availability.
Not yet determined.
Getting started
Nothing during alpha. Early access is sandbox access with no billing. The pricing page has the $25 per million tokens rate planned for general availability.
Through the early access form. Every application is read before anyone is let in, and onboarding happens in batches.
People test-driving the platform, people interested in platform engineering or agentic workflows, and agencies.
Your business documents. The AI grounds what it produces in the documents, product notes and content you give it, so upload those first.
Website content, yes, by pointing the platform at your existing site so it reads the pages. CRM contact and email list imports are not available yet.
Yes. Custom domains with SSL are supported today.
During alpha, by email or on the Discord server.
The rest is quicker to try than to read
Ask for early access, upload a document and read the first draft. Anything still open goes to the contact page.