Lindy

What it is

Lindy is a no-code platform for building always-on AI agents that run over your email, calendar, and SaaS tools. You describe what you want in plain language, pick a trigger (an email lands, a form is submitted, a meeting ends, a row changes), and Lindy assembles an agent that takes the steps and calls the tools you allow. It is built primarily on Claude — Claude Sonnet 4.5 is the default model, with GPT-5 and Gemini also selectable — and connects to more than a thousand apps. The thing that separates it from a chat window is that agents fire on events without you in the loop: this is automation that reasons, not a prompt you babysit.

Lindy 3.0 (August 2025) added the pieces that make it usable for real ops work: an Agent Builder for going from prompt to a running agent in minutes, Autopilot (a cloud browser the agent operates like a human — click, fill, scroll, navigate — for the long tail of apps with no API), and team accounts. The company was founded in 2022 by Flo Crivello (early Uber PM, previously founder of Teamflow) and has raised roughly $50M to date.

Why it shows up across RevOps, Legal, and Recruiting

The same builder serves all three because the job is identical — an agent watching an inbox or a queue and acting:

• RevOps runs an inbound agent that reads a demo request, enriches the company, drafts a routed reply, and books the meeting; or a meeting agent that turns the call recording into a CRM update and a follow-up draft.
• Legal ops points an agent at a shared intake inbox to triage NDA requests, tag them by clause risk, and draft the first acknowledgment — with the redline left to a human.
• Recruiting runs an agent that screens inbound applicants against a scorecard, schedules first-round calls against the recruiter’s calendar, and nudges candidates who go quiet.

Pricing

Lindy 3.0 prices on tiered monthly plans with a usage allowance, not a per-seat flat fee:

• Plus — $49.99/month, standard usage allowance, up to 2 inboxes. The entry point and the right place to pilot.
• Pro — $99.99/month, ~3× the Plus usage allowance, up to 3 inboxes.
• Max — $199.99/month, ~7× the Plus allowance, up to 5 inboxes.
• Enterprise — custom, for teams that need SSO, shared accounts, and controls.

All individual plans come with a 7-day free trial, no card required. The headline number is not the whole bill: usage is metered against your allowance, so multi-step agents, Autopilot browser runs, and especially voice (phone numbers and per-minute call time are billed on top) consume far faster than a “send an email” mental model predicts. Budget the allowance you’ll actually burn, not the plan price.

Best for

• Founders, ops generalists, and lean RevOps / Legal / Recruiting teams (1–50 people) who want always-on agents over email, calendar, and their SaaS without writing code or standing up infrastructure.
• Teams whose highest-value automations need judgment mid-flow (read this, decide, then act) rather than a fixed if-this-then-that path.

Alternatives and when to pick them

• Zapier — pick when the work is deterministic plumbing (move data A→B on a trigger) across the widest integration catalog. Zapier owns the most market share in workflow automation; choose it when you don’t need an agent to reason, just to connect.
• n8n — pick when you want self-hostable, open-source workflow automation your engineers can own and extend. The fastest-rising option for technical teams that need data residency or want to avoid per-task metering.
• Dust — pick when the priority is governed agents over permissioned company data (audit trail, dual-layer permissions, EU residency) for an org of 20–500 rather than a personal assistant. Dust leads on governance; Lindy leads on speed-to-first-agent and ease.
• Gumloop — the fastest-growing AI-native entrant: pick it when you want a visual drag-and-drop canvas for document-heavy pipelines and prefer cheaper credit economics to Lindy’s chat-first build.

Watch-outs

• The usage allowance is the real budget, not the plan price. Computer-use runs and voice minutes burn the allowance fast, and voice is billed separately on top. Guard: run the 7-day trial on your single highest-volume workflow, read consumption off the dashboard, and size the tier from that — don’t pick a plan off the headline price.
• Autonomy without a gate sends wrong actions at machine speed. An agent that emails prospects, messages candidates, or drives a browser will act on bad enrichment or a misread instruction faster than you can catch it. Guard: keep send/write actions in draft-or-notify mode until the agent has earned trust on a sample, then promote individual steps to fully automatic one at a time.
• It is an agent builder, not a system of record or a governance platform. Lindy is lighter on enterprise permissions, audit, and data-residency controls than tools built for that. Guard: before routing regulated or permissioned data through an agent, confirm the Enterprise plan’s controls cover your requirement, or keep that data path on a tool like Dust.