Background
In early 2023 chat assistants built on large language models were attracting wide attention, and several companies were preparing their own. Anthropic is an AI company, and Claude was its assistant. An AI assistant is a chat-style system that takes natural-language requests and replies in conversation, handling tasks such as drafting text, answering questions and writing code. Before a broad release, companies often run a closed alpha, an early test in which only invited partners and users can try a product. Feedback from that stage shapes what gets released.
What happened
On 14 March 2023 Anthropic made Claude, which it called a next-generation AI assistant, available more broadly after a closed alpha.[1] Partners including Notion, Quora and DuckDuckGo had tested the system during that alpha over the previous few months.[1]
Claude was offered in two versions: Claude, a high-performance model, and Claude Instant, a lighter, less expensive and much faster option.[1] It was available through a chat interface and through an API in Anthropic's developer console, and organisations could request access.[1]
Key details
Anthropic said Claude could handle summarisation, search, creative and collaborative writing, question answering and coding.[1] The company said the system was based on its research into training helpful, honest and harmless AI systems.[1] It also said Claude could take direction on personality, tone and behaviour.[1]
Launch partners included Quora, which used Claude in its Poe chat app; Notion, through Notion AI; DuckDuckGo, through DuckAssist; the legal-technology firm Robin AI; Juni Learning, through a Discord tutoring bot; and AssemblyAI.[1] These partnerships show one route by which Claude reached end users: inside other companies' products. The partners were varied, from a consumer question-and-answer site to a legal-technology firm and a Discord tutoring bot, so the same underlying system appeared in very different settings.[1]
The faster, lighter Claude Instant was positioned for applications where response speed and cost matter more than the largest model's capabilities. Offering both versions let developers choose between the two according to the job.[1] Because the product was reachable through both a chat window and an API, everyday users and developers reached the same underlying system in different ways.[1]
Impact
On 11 July 2023 Anthropic released Claude 2 and opened a public beta chat website, claude.ai, to anyone in the US and UK.[2] That step moved Claude from invited organisations and partner integrations to a service that individuals could sign up for directly.[2] Launch partners and the public website together show two paths of access, one through products built by others and one through Anthropic's own interface.
Anthropic's stated design goal, helpful, honest and harmless behaviour, is a claim about intent. How closely a given model meets that goal is something users and outside researchers judge through its actual responses.
Sources
- 1Introducing ClaudeAnthropic
- 2Claude 2Anthropic
Cite this page: Anthropic introduces Claude. TellWorld, https://www.tellworld.net/news/claude-2023. Updated 2026-11-07