Lex Review, Pricing & Alternatives (2026)

Lex is an AI-focused word processor founded by Nathan Baschez and spun out of Every. TechCrunch reported a $2.75 million seed round led by True Ventures on August 23, 2023, after Lex attracted about 25,000 sign-ups during its first day. Lex’s 2024 annual letter reported 47,000 users making 15 million edits across 400,000 documents, alongside more than 14 trillion language-model tokens processed. The product targets substantial writing and revision rather than standalone text generation. It combines a document editor with AI assistance, comments, shared documents, folders, and model selection. Its current pricing page says Lex is no longer accepting new customers, which materially changes its usefulness for new buyers while existing accounts continue working. Lex remains positioned around AI-assisted writing inside an integrated word-processing workspace.

Lex’s workflow keeps drafting, reference material, conversation, and revision in one browser-based writing environment. Users can write documents, collaborate through comments and sharing, open Ask Lex, attach files or PDFs, and request ideas, counterarguments, rewrites, summaries, or feedback on selected text. Context tags collect reusable background information from files, URLs, Lex documents, and folders, then let users add that material to chats when needed. The 2024 annual letter describes context tags as buckets of reusable information that can be shared across teams, allowing one person to maintain reference material for others. Checks provide inline suggestions, including a Brevity Check, while folders organize documents and shared work. The workflow is aimed at writers who want AI to support judgment, editing, research, and iteration rather than replace authorship.

Lex’s model selector supports models from OpenAI, Anthropic, TogetherAI, Perplexity, Google, Mistral, and Meta according to its model guide, with availability varying by account and tier. The guide dated April 24, 2025 lists GPT-4.1 as the Pro recommendation and GPT-4.1 mini as the free recommendation, while Lex’s token guide describes context limits and usage costs by model. Lex announced Lex V2 on April 3, 2025, describing five major releases and showing a redesigned start page; the same announcement references voice-to-text in Ask Lex. The pricing page says Pro includes premium models, early feature access, the iOS app, and priority email support. Lex also says it uses enterprise APIs and does not train language models on user writing.

Lex published a Free tier at $0 with free models and limits on Ask Lex messages, Checks, and custom prompts, although the current page does not publish a complete numerical quota. Lex Pro was published at $18 per month when paid monthly or $12 per month when billed annually. Pro includes premium models, early feature access, priority support, and high usage limits; Lex says overage credits may be available at cost. Lex Teams was published at $8 per user per month when paid monthly or $5 per user per month when billed annually. Teams adds unlimited group folders, shared templates, shared prompts, and team collaboration. These prices are for existing accounts because Lex says it is no longer accepting new customers.

Lex’s main caveat is availability: new customers cannot subscribe according to the current pricing page, while existing accounts continue working. The free plan’s complete numerical limits are not publicly stated, and Pro usage limits are described as generous rather than given as a fixed quota. Lex’s privacy notice says it collects documents, comments, chat inputs, assistant outputs, account information, usage information, and related metadata. Its terms grant Lex a license to operate the service and provide personalized suggestions, while stating that uploaded content is not used to train its models. Google Docs, ChatGPT, and Grammarly are named direct competitors, with Notion AI, Hemingway Editor, and Writer also relevant alternatives. Existing Lex users should pick it when integrated document editing and reusable context matter most.