Lindy
An AI agent platform for lightweight business tasks, communication and integrations.
Where it fits
An AI agent platform for lightweight business tasks, communication and integrations.
Price and access
Check the current plan before paying
A free option is available for trying real tasks. Paid plans or usage can add higher limits or extra features.
- Plus$29.99 / monthAs of 2026-09-05 · source ↗
- Pro$99.99 / monthAs of 2026-09-05 · source ↗
- Max$199.99 / monthAs of 2026-09-05 · source ↗
Prices are collected from the official site and can change. Always confirm the current plan, taxes and regional availability before paying.
What it helps with and what to check
Good tasks to start with
- Lightweight business agents
- Trying a bounded multi-app workflow
Keep an eye on
- Review every connected-account permission
- Keep approval before messages or external actions
How to start
Tick each step as you work. These checks stay on this page only; they do not verify results in the external tool.
0 of 5 steps checked
Check the workspace before trying an action
This exercise uses Lindy’s current Teammate documentation and fictional requests. In Integrations, inspect both personal and team connections. Use only a practice workspace with no connected work accounts and no active routines. Do not disconnect a shared connection just for this exercise. If you cannot verify that setup, do the policy worksheet below instead. A request saying “do not act” is not a permission barrier.
Ask for a permission plan, not a real task
In Home or Chat, ask: “Use only these fictional cases and the official integration documentation at https://docs.lindy.ai/integrations/overview. Do not use integrations or create routines. Explain whether Require approval covers: A, writing through an integration in a shared Slack thread; B, the same write in web chat; C, reading through an integration in that Slack thread. Give the documented scope and unresolved questions. Return text here only.” If access requires connecting a real account, stop and use the worksheet.
Compare the answer with the actual boundary
The documented answer is: A asks before writing when Require approval is set; B is outside these guardrails; C is not guarded because reads are never guarded. The same limitation applies to Slack DMs, iMessage and SMS: the documented guardrails do not cover those surfaces. Open the official page and check these statements yourself. This is a documentation-based answer key, not an observed Lindy run or a security test of your workspace.
Fix a promise that is broader than the setting
If the answer says “every action needs approval”, ask it to replace that claim with the exact supported scope. Also distinguish Approve, which permits one requested write, from Always allow, which changes the rule going forward. The docs say an owner/admin can change it workspace-wide and a member can change it for themselves. Do not click either control in this exercise. If the answer promises separate rules for two accounts under one integration, correct that too: the documented guardrail covers all accounts under that integration.
Write the requirements before making it recurring
Use the requirements worksheet below to record the surface, integration, account scope, allowed reads, proposed writes, approver and stop condition. Keep unanswered items marked unknown. Then use the policy Task to review them with the responsible person. Routines add a trigger, prompt and destination; they are a later decision, not proof that approval applies everywhere. Before a real rollout, test the permitted and denied cases in your actual setup and record what happened.
Platforms
Privacy and human checks
Before you upload
- Remove customer names, passwords, identifying details and company secrets.
- Read the provider’s current retention, training and workspace settings.
Before you use the result
- Verify facts, sources, calculations and code against the original material.
- Have a person review important content or actions before publishing, sending or running them.
Related tasks and guides
Prepare inputs or review results in these AhaDo workflows. Copy permitted material yourself when using an external tool.
Alternatives
A no-code automation platform with AI workflow and agent features across many software services.
View comparison option MakeA visual no-code platform for building multi-step automations and data flows.
View comparison option ManusAn autonomous task agent for multi-step research and execution workflows.
View comparison optionOfficial source and research status
Guide published · Official Lindy documentation read on 7 September 2026. This permission-planning exercise has a documentation-based answer key; account access, guardrail enforcement and provider execution were not tested.
Listed means a published directory entry. Guide published means it also includes dedicated start steps and sourced prices. Neither status means real-user results have been verified.
- Official product source
- Open source ↗
- Last researched
- 2026-09-07
- Real-user evidence
- Not collected yet
Report a correction
Has a tool's information, price or signup changed? Tell us and help the next beginner.
