Kavela vs Make

1. Building an automation

How Make does it

The standard way to build in Make is the Scenario Builder: you drag modules onto the canvas, connect them, map fields between them and add routers and filters ¹. One independent review says Make demands more upfront learning than Zapier and estimates two to three hours to feel comfortable, against 30 minutes for Zapier ¹⁰.

Maia builds and edits scenarios from a prompt or an uploaded sketch; each change appears on the canvas as she makes it, and you can go back to an earlier version ^{2, 4}. Make's system card for Maia sets clear limits: 'Maia is unable to take high-impact actions, such as running, activating, or deactivating scenarios. She asks you to perform these actions yourself,' and you must check a scenario's configuration before running or activating it ⁴. Maia cannot see your other scenarios, so she cannot reference them or call subscenarios ².

Make also reaches outside its editor. Of the official ChatGPT plugin, released on 9 September 2026, Make says: 'Describe the work you need in plain language. Make connects the required apps and sets the automation to run on a schedule or a trigger.' It uses your existing Make connections and works on all Make plans with a ChatGPT account ³. Make is also a built-in connector in Claude for running scenarios, and Make Skills, in open beta, teach Claude to build scenarios through the Make MCP server ^{11, 12}.

How Kavela does it

In Kavela you start in a conversation. You tell Kav what you want, in the workspace chat or in Telegram, the way you would brief a colleague: every Friday, pull the week's sign-ups, write a short summary and put it in the team's shared folder. Kav sets the automation up, connects the apps it needs or asks you to through an Approvals card, sets the trigger and runs it.

Make's ChatGPT plugin shows that describing an automation in a chat and having it set to run is no longer unusual ³. The real difference is how far the agent goes. The plugin works from the Make connections you already have ³. Kav also sets up the connections that are missing, asking you only through a card when it needs you, runs the work on your own model plan, and hands the steps that need judgement to an agent.

The canvas is a good place to inspect a scenario; describing the job should be enough to get one running.

2. Triggers and schedules

How Make does it

A Make scenario starts from an instant trigger, a webhook that runs 'immediately after the webhook URL receives a request', or from a polling trigger that checks an app on a schedule; webhook processing can itself be scheduled in batches ⁵. Scheduled runs can be as frequent as every 15 minutes on Free and every minute on paid plans, and a run on Free lasts at most five minutes ¹.

For email, a custom mailhook gives a scenario a Make address you forward mail to; Make's guide sets it up with a Gmail forwarding rule and a filter, and lists two email accounts as a requirement ¹³. Make's guides also cover agents triggered from Slack, Telegram, email, forms and webhooks, and note that other agents can run on a schedule ¹⁴.

How Kavela does it

Kavela automations run on a schedule or on an event, such as a webhook or a merged pull request. You say when, in plain words, and Kav sets the trigger. Each automation can be limited to an active window with a start and an end date, so work tied to a launch or a campaign stops when it ends.

Each run happens in the agent's own cloud computer, which starts when the work starts and sleeps when it stops, and work keeps going after you close the tab. The workspace's files persist between runs, so tomorrow's run starts from today's output. There is no cap on how many are switched on at once.

Saying when a job should run, and when it should stop, is a sentence for you and a setting for the agent.

3. Your own model plan

How Make does it

AI in Make is paid in credits. With Make's own AI Provider, available on every plan, an AI step costs 1 credit per operation plus credits based on AI tokens, and since 25 August 2026 input and output tokens are priced separately ^{6, 15}. With your own provider connection, such as Anthropic Claude, OpenAI or Gemini, it costs 1 credit per operation and the token costs go to your provider; Make's help pages say custom provider connections are on paid plans ^{6, 7, 8}. Maia's own usage also uses credits on paid plans, after some free introductory messages ².

The docs describe no way to run AI steps on a ChatGPT or Claude consumer subscription; the own-key option is a provider connection ^{6, 7}. When you build through the ChatGPT plugin, the conversation runs in ChatGPT, and the scenario it sets up runs in Make, where each module action counts as a credit ^{1, 3}.

How Kavela does it

Each Kavela agent runs on the plan you choose for it. With your own Claude plan, the unmodified Claude Code runs inside the agent's own cloud computer and you sign in to it yourself, which is the form Anthropic's terms describe for platforms ¹⁶. With a ChatGPT plan, the agent runs Codex instead, and anyone with neither uses Kavela's included Kav AI usage.

The choice is per agent, so one automation can run on Claude and another on Codex. The AI work in your automations draws on subscriptions you already pay for, with no provider connection or API key to set up.

Credits are counted per module and per token; running AI work on a plan you already pay for leaves one less meter to watch.

4. Work that needs judgement

How Make does it

Make AI Agent (New), an open beta whose functionality and pricing may change, puts an agent into a scenario through a 'Run an agent' module with an instructions field ^{7, 8}. The agent can call modules, scenarios and MCP tools, pass work to sub-agents (added on 27 July 2026), use knowledge files and switch to a fallback model connection if the first one fails, and it shows its reasoning step by step ^{7, 8, 17}.

For inbox work, Make documents email- and mailhook-triggered agents ^{13, 14}. For the web it offers AI Web Search, in beta ¹, and it deprecated its Google Chrome app and browser extension on 31 August 2026 ¹⁷. The agent app's page does not mention approvals ⁸, and the docs describe no agent that signs in to a web app and tests it in a browser, and no lead agent that plans work across a team ^{7, 8}.

How Kavela does it

In Kavela, a step that needs judgement is work an agent does, with the tools a person would use. Each agent has its own cloud computer with a shell, files, git and a browser. It can read a customer's email, work out what they are asking and write the reply, or sign in to your app in its browser, try the new feature and report what broke.

The work belongs to a standing team. Kav, the lead agent, plans the job, hands each part to the right worker agent, follows up and brings back the result, and every task sits on one calendar with its start time, due time and live state. If no worker fits, you describe the job and Kav builds one that follows that process.

A judgement call is hard to draw as branches on a canvas, and easy to describe to an agent.

5. Connecting your tools

How Make does it

Make connects to web services through OAuth 2.0 connections, keys and certificates, and connecting Google services can mean creating a custom OAuth client in Google Cloud ¹⁸. Maia shows a connection card in the chat and you enter the credentials in a dialog outside it; in Make's words, 'Maia never has access to your connections or keys' ⁴. The ChatGPT plugin reuses the connections you already have in Make ³.

Credential requests, added in February 2026, let you send a link so someone else authorizes a connection for you; anyone can receive one, but sending is limited to official Make Partners and Enterprise customers on request ¹⁹. Make agents can add MCP tools, and Make runs its own MCP server ^{8, 20}. A mailhook address receives mail you forward to a scenario, and the docs do not describe an agent using it to sign up for services ¹³.

How Kavela does it

Kavela connections belong to the workspace: connect a service once and every agent can use it. A service can be connected by OAuth sign-in, as an MCP server, with an API key, with custom secrets, or by signing in to it in a live browser you control. Passwords and keys go into the card, never through the model.

You can also leave it to the agent. With Let Kav set it up, the agent reads the service's documentation, works out the right way to connect it and does the setup. If there is no account yet, it creates one, using the workspace's own receive-only inbox to read the verification email.

Make keeps credentials away from Maia by design; Kavela keeps them away from the model too, and still lets the agent do the setup.

6. What it needs from you, and what it delivers

How Make does it

Make's own Human in the Loop app, with modules to create and cancel review requests, list reviews and watch completed ones, is in closed beta, available only to invited customers on the Enterprise plan ⁹. A separate third-party HITL app is listed in Make's app directory and needs its own account and API key ²¹.

What a scenario produces shows in its run history and logs, searchable across scenarios and runs since August 2026 ¹⁷, and lands in whatever app the scenario sends it to. When you build through the ChatGPT plugin, results come back in ChatGPT as tables and status badges ³. The docs describe no single inbox that gathers approvals from every scenario ⁹.

How Kavela does it

Kavela collects everything your agents need into one Approvals inbox for the whole workspace. Each request is a typed card you answer where it is: connect an app, sign in to a site in a live browser, approve an action before it goes out, choose between options, or let an agent know the email it was waiting for has arrived. Other work carries on while a card waits.

What agents produce arrives as deliverables attached to the task that made them, from documents and reports to sites and files, listed under Delivered and opened in one click.

When every request lands in one inbox, nothing waits on a channel you forgot to check.