Delegate email, meeting follow-ups and team tasks to an AI assistant with connected tools, reusable skills and scheduled routines. Lindy works in Slack, web chat and phone messaging.
Used this tool? Rate it
Used this tool? Rate it
Lindy is an AI assistant for connected tools, inboxes and meetings, positioned as an AI employee that learns how you work and completes delegated tasks. Its capabilities depend on connected accounts, accessible sources and permissions.
A shared assistant works visibly in Slack channels; a private assistant handles personal inboxes, meetings and scheduling in direct messages. Both spend the workspace's credit pool.
Email automation separates labeling, drafting and unanswered-message follow-ups. Labeling sorts incoming mail and can archive it, but needs at least one label to work. This built-in routine lacks custom trigger, schedule, delivery, Smart Filter and run-history settings.
Email Drafting saves replies for review, editing and sending in a connected Gmail or Outlook inbox. Tone, length, signature and skip or escalation rules are set per inbox. The company says Lindy also learns from edits. Drafting itself does not send the email.
Lindy records Google Meet, Zoom and Microsoft Teams calls, captures notes and action items, and answers questions across accessible recordings or folders. Folders can be private, shared with selected people or team-wide. Connected apps enable follow-ups such as HubSpot updates, Notion handoffs and tasks from action items.
Watch and Learn turns a demonstrated process into a reusable skill and a routine.
Demonstrations accept MP4, MOV, WebM and M4V video, plus PDF, TXT and Markdown instructions. Limits are up to 2 GB per file, one hour across all videos including a new recording, and 4,000 note characters. At least one recording or file is required; analysis uses credits.
A routine dry run uses real triggers, reads and judgment but suppresses sends, posts and writes. It previews classifications and drafts, while leaving actual delivery untested.
As captured on October 25, 2026, a software-review seller profile showed 4.9/5 from 171 Lindy reviews; a company-review profile showed 1.9/5 from 41 reviews. The latter warns its sample may not be representative and includes an unrelated coffee-shop complaint. Different samples and scopes prevent treating these as a common product-performance benchmark.
The public pricing page captured on October 25, 2026 shows a Free entry with $50 in credits lasting 7 days and no credit card required. Team starts at $49.99 per active user per month for 5,000 monthly credits, pooled across the workspace. Enterprise uses custom credits and billing, with features including SSO, SCIM, audit logs, dedicated support and a signed BAA for HIPAA-related use.
The captured Team selector lists these monthly credit choices, without promising a fixed number of finished jobs:
| Monthly credits | Displayed monthly price |
|---|---|
| 5,000 | $49.99 |
| 6,000 | $59.99 |
| 7,000 | $69.99 |
| 8,000 | $79.99 |
| 15,000 | $99.99 |
| 20,000 | $149.99 |
| 35,000 | $199.99 |
| 45,000 | $299.99 |
| 70,000 | $399.98 |
| 105,000 | $599.97 |
| 175,000 | $999.95 |
| 350,000 | $1,999.90 |
| 700,000 | $3,999.80 |
| 1,750,000 | $9,999.50 |
Illustrative usage bands are 2–250 credits for everyday asks, 250–1,000 for deep work, and 1,000–2,500 for large builds such as dashboards and decks. They are not fixed message prices. Most major models are described as available, with selection per task.
Monthly plan credits reset at the billing-cycle start and do not roll over. Purchased top-up credits carry over and do not expire; the public rate is $10 per 1,000 credits. Pricing documentation says upgrades take effect immediately, downgrades at the end of the current term, and credit and seat allocation changes immediately. It names Plus, Pro and Max, while the public page groups the selectable credit offering as Team; this capture does not establish a full price mapping between those labels.
Base-plan usage is described as $0.01 per credit. Shared channel work and private assistant work draw from the same pool. Phone calls start at 25 credits per minute, with a separate monthly fee for the phone number. An admin can cap a member's usage, but an already-running task can finish beyond the cap before further credit-using actions pause. The cap is not an exact cutoff inside a running task.
Lindy says source breadth, routine frequency, thread history and model choice affect cost. Long threads re-read earlier messages; routines spend when they run. Heavy tasks can pause for a continue/stop choice in the originating thread or DM. Stopping leaves the thread open without further task credits.
New teammates entering through Slack get a one-time 7-day trial, and admins receive a notice with a removal link. Returning teammates do not receive another trial. A paid seat stays billed until removed by an admin and remains available through the paid cycle without mid-cycle proration. The public pricing FAQ separately says direct signups are billed right away, alongside its no-card free-credit offer; those entry conditions are presented separately rather than as a universal free week for every signup.
Cancellation takes effect at the end of the current paid term, and the terms state that purchases are non-refundable. The terms also leave included allowances to the selected plan or order form and permit enabled overages where available. This is broader than the pricing FAQ's pause-when-empty description: the terms distinguish accounts with overages enabled from those without them. No separate trial-refund window was identified in the checked current pricing, billing documentation and terms. A conditional mobile-license clause allows an App Distributor to refund an app purchase under its own policies after a warranty failure; it does not establish a web-subscription refund window.
Make addresses visual workflow automation. Its maker describes building agents and workflows by prompt or drag-and-drop, connecting apps and data, and adding custom API integrations, with a free plan without a time limit. Visual process construction is the comparison point against Lindy's personal or shared assistant.
n8n addresses hosting control: deployment on the operator's infrastructure, on-premises or in a private cloud. Self-hosting without a license key runs the free Community edition; Business and Enterprise licenses enable other editions. This offers deployment ownership and edition choice, without establishing parity with Lindy's assistant experience.
These are respective vendor capability statements, not comparative reliability tests.
The public pricing FAQ says outside-impact actions wait for approval, excluding approved-source reads. Integration documentation states narrower rules: Always allow writes without asking, Require approval asks in Slack, and Don't offer prevents writes. Those guardrails cover shared Slack threads only, excluding web chat, Slack DMs, iMessage and SMS. The pages state different scopes; the broad FAQ is not a description of every documented path.
Shared integrations use the connected account's permissions for everyone. A guardrail covers every account under an integration, preventing separate rules for work and personal Gmail. Slack requesters, owners or admins can approve. Approve permits one action; Always allow changes the rule workspace-wide for an owner/admin or personally for a member.
Saved credentials support direct API calls through shell/code, start personal until shared and cannot be read back. The credentials table marks guardrails as applicable to integrations, not credentials. MCP separately offers Always allow or Ask for approval, with personal choices overriding workspace defaults.
CC-based scheduling requires both Lindy on CC and scheduling intent. It books agreed times automatically with invites, or replies with three available slots. This documented trigger can lead to external action beyond saving a draft.
Phone access requires a paid plan and may vary by account. Approved calls can navigate menus, hold, leave voicemail and report back; outside callers cannot access private inboxes, files, memory or conversations. Personal routines can call within 8 AM–9 PM recipient-local hours and line limits. Script routines are personal-only and lack connected Gmail, Slack or Calendar credentials.
Phone terms make users responsible for consent, message rights, required opt-outs and do-not-call checks, including prior express written consent for telemarketing/AI voice where legally required. They also exclude contacting or supplying non-US numbers. These contractual requirements do not automatically authorize every use.
Inbox limits conflict: public pricing says two to five depending on plan; Email Drafting documentation says as many as desired across Google and Outlook. The checked pages do not establish a single common maximum.
Lindy says customer data is encrypted in transit and at rest, is never sold, and is never used to train models; Enterprise adds SSO, SCIM and HIPAA with a signed BAA. These are vendor statements with plan qualifications.
The privacy policy names Anthropic, OpenAI, Perplexity AI, Exa Labs, ElevenLabs and Deepgram as AI providers. Google API data is excluded from AI training and evaluation. The terms say Lindy and its AI providers do not train on Contributions; this is not a claim that data never reaches service providers.
Retention lasts as necessary, with legal exceptions. Without an ongoing legitimate need, data is deleted or anonymized; backups may be securely isolated until deletion is possible. Information may remain after termination for fraud prevention, troubleshooting, investigations, enforcement or legal requirements. No fixed deletion deadline is promised.
Users retain Contributions ownership while granting a limited service-operation license and an irrevocable perpetual analytics license; public Contributions may be used lawfully. A separate blanket grant of exclusive commercial rights to every AI output was not found in the checked terms and privacy policy.
No. Automatic drafts are created without the extension. The extension adds Gmail controls for refining and sending them inside the inbox; its Outlook support is described as coming soon.
A skill describes how to perform a procedure. A routine determines when work runs and where its result goes. A demonstrated process can become both.
Monthly plan credits expire at the billing-cycle reset. Purchased top-up credits remain until used, so the two balances have different rollover rules.
