Introducing Lindy 2.0 - The FIRST True AI-First Automation Platform

Lindy 2.0 positions itself as a true automation platform rather than an “AI assistant” that only responds to prompts. The core shift: Lindy can run multi-step workflows that connect to real services—like Google Calendar and Google Docs—so tasks can execute automatically once a trigger fires. That matters because it moves AI from drafting text to performing actions, with built-in control points that decide when an agent should continue, stop, or hand off to the next step.

The walkthrough starts by contrasting ChatGPT-style usage with what Lindy enables. ChatGPT-like tools require the user to initiate every action and provide the right inputs; they don’t autonomously manage external systems. Lindy’s approach centers on “Lindies” built through a flow editor. Users begin with a message node that kicks off a bot, then route execution into one or more “AI agents.” Each agent runs under a defined prompt and model choice, and it can be configured with “skills” that grant access to specific actions—such as creating, listing, updating, and deleting Google Calendar events.

A calendar example makes the difference concrete. A custom “calendar management bot” greets the user, checks Google Calendar for conflicts, proposes the best available times, and then creates a new event with details and a link. The workflow also demonstrates recurring automation: a trigger set for every day at 8:00 a.m. launches a separate agent that summarizes the day’s schedule and upcoming events. After the calendar check completes, “jump conditions” determine when the flow should exit the agent and move to the next task—such as sending an email summary—so the system doesn’t stumble between steps.

Lindy 2.0 then expands beyond calendar management into information gathering and document automation. Using a “web monitoring” template, the creator builds a news gatherer that extracts text from web pages and PDFs, updates a Google Doc with the latest AI news, and can optionally record changes to Slack. The example evolves into a multi-source pipeline: it searches the web (including Bing), extracts from an AI news site, and even searches YouTube for AI news results, then writes the aggregated findings into the same Google Doc with correct dates and formatting. The flow editor also supports conditional logic—only continuing when new content is detected since the last check.

Under the hood, Lindy offers model selection (including OpenAI and Claude options such as GPT-4 Turbo, GPT-4o mini, Claude 3 Haiku, Claude 3 Sonnet, and Claude 3 Opus) and a large menu of integrations and actions. The transcript highlights connections to Google Workspace (Calendar, Docs, Sheets, Tasks), Gmail, Slack, HubSpot, AWS-related tools, HTTP/custom API calls, webhooks, and more—plus “premium actions” gated behind higher tiers.

Finally, Lindy 2.0 is framed as credit-based automation: running AI tasks costs credits, with a free intro month (up to 400 credits) and paid plans that replenish monthly. The creator emphasizes that Lindy is early access, with new flow-chart features and ongoing improvements, and notes cross-device access via iOS and desktop apps for Mac and Windows. The overall message is that Lindy 2.0 aims to make AI-driven apps that can do real work—without constant oversight—by combining triggers, agent logic, and direct service permissions in one workflow builder.