Gentlebot lets you spin up one agent or a full team. They watch inboxes and websites, run on schedules and triggers, send emails and SMS, build files, reports, spreadsheets, and live web apps.
No flowcharts, no brittle workflow builder. Give each agent a role, tools, connections, and triggers — then watch the work run live.
Name the role, describe the job, and grant only the tools and connections it needs: Gmail, Slack, Stripe, Sheets, browser, APIs, and more.
Start it by chat, @mention, schedule, webhook, inbound email, or SMS. It can keep working quietly until something needs your attention.
Receive the actual output: a reply, a spreadsheet, a PDF, screenshots, a Slack update, a file, or a hosted web app with a live view.
Gentlebot is built for reusable agents, not one-off prompts. Create a sales agent, an inbox agent, a research agent, and an ops agent — then run them separately or compose them inside shared workspace threads.
Agents can start on a schedule or wake up when something happens in your tools.
The result is not another chat transcript. It is something your team can open, send, share, or build on.
From following up with leads to preparing reports, create agents for the jobs that keep coming back.
Turn attention, leads, and research into pipeline-ready work.
Move recurring coordination, cleanup, and reporting off the team.
Give one person a small team of always-on assistants.
Coordinate specialized agents across teams, tools, and approvals.
Native connections for the tools you already use, plus HTTP APIs and MCP servers for everything else. Credentials stay encrypted and never reach the model.
Bring your tools. If it has an API, inbox, webhook, or MCP server, an agent can usually work with it.
Create a small team of agents for free. Upgrade for more work, more agents, and more tasks running at once — with no surprise overage bills.
Credits cover AI processing, browser time, and sandbox compute. Monthly credits reset on your billing date. Pro credit packs expire 30 days after purchase, and queued tasks use no credits. See how credits work →
How agents, credits, concurrent work, and limits fit together.
An agent is an AI worker with its own instructions, tools, connections, files, memory, and triggers. Create separate agents for support, research, reporting, operations, or any other job.
Yes. Schedules, webhooks, APIs, and incoming email can start an agent while you are away. “Always available” does not mean an expensive browser or sandbox stays running while idle.
Credits pay for variable work: AI processing, active browser time, sandbox compute, and paid tool calls. More complex or longer tasks consume more credits.
Yes. Credits belong to the workspace and are shared across its agents. You can also give individual agents their own budgets so one automation cannot consume everything.
It is an agent task currently executing. Free runs one task at a time and Pro runs up to five. When every slot is occupied, new tasks wait safely in a queue and consume no credits while waiting.
No. A Pro workspace can configure 25 agents while up to five execute simultaneously. Dormant agents cost almost nothing; the work they perform consumes credits.
Browser and sandbox time is metered only while those resources are active. Using both costs about 1.5 infrastructure credits per minute, plus AI usage. Sleeping resources stop their duration meter.
Yes. Free runs can last up to 10 minutes and Pro runs up to 30 minutes. These limits stop accidental loops and surprise usage. Custom plans can use tailored limits.
New work pauses and active work stops safely. Pro customers can purchase a prepaid pack without changing plans; Gentlebot never switches on postpaid overages by surprise.
Purchased credits expire 30 days after purchase. Gentlebot always uses the credits that expire soonest first.
No. Monthly plan credits reset on your billing date. Purchased credits have their own 30-day validity window.
Yes. Every active run has a stop control. You are charged only for AI and active resources used before the run stops.
Provider usage already incurred may be charged. Credits caused by an internal Gentlebot platform failure are returned automatically after usage is reconciled.
Uploads, generated documents, artifacts, browser downloads, and retained run outputs count toward workspace storage. Small internal configuration metadata does not.
Yes. Agents can use supported integrations, HTTP APIs, webhooks, and MCP servers. A third-party service may charge its own fees independently of Gentlebot.
Custom plans offer higher agent, credit, storage, concurrency, and duration limits, plus team access, governance, onboarding, priority support, and optional service-level agreements.
Create your first always-on agent, connect the tools it needs, and let it start delivering finished work.