Nekton
Nekton turns plain-English descriptions of everyday tasks into working cloud automations: the AI writes the code, runs it on demand or on a schedule, and connects to thousands of apps. A permanent free plan is included.
What is Nekton?
Nekton is a workflow automation service built around one idea: instead of assembling blocks on a canvas, you write down what you want done, step by step, in ordinary English. Where a step can be automated, the AI writes the code for it and executes it in the cloud. The homepage compresses the loop into three words, Document. Run. Automate with AI., and the vendor states the engine is powered by OpenAI and ChatGPT.
What makes the approach workable is that automation is not all-or-nothing. A workflow can hold manual steps alongside automated ones, so a team can write a procedure down first, run it by hand, and automate the painful parts as it goes. Workflows can collect input from people through forms, whose fields can be pre-filled from URL parameters, and they can be triggered on demand, on a schedule, or through a shared link that anyone can run without creating an account.
The integration surface carries the other half of the pitch. Nekton publishes a category index spanning roughly eighty-five areas, from CRM, e-commerce and databases to team chat, storage, transcription and AI tools, and its sitemap lists more than 5,600 individual integration pages, one per application. Anything reachable over an HTTP API is in scope, and developers who hit the limits of the generated code can write their own steps in TypeScript.
Beyond the core product, Nekton sells an API of its own. An HTTP interface drives the service from another application, or an embeddable UI lets your own users describe and build automations inside your product. Access is granted on request rather than self-service.
The vendor is unusually candid about the boundaries. A section of the homepage lists what the tool handles badly: complex or vaguely stated goals, automation that must run locally or reach systems closed to the internet, and large data volumes. The published use cases stay deliberately modest, covering employee onboarding, incident handling, hardware requests, release checklists, GitHub reporting, keyword monitoring, customer surveys and lead follow-up. One caveat belongs in any description: nothing on the site has been publicly updated since October 2023, although the service still responds.
What it does
- Describe a task step by step in plain English and let the AI write the automation code
- Run the resulting workflow on demand, on a schedule, or through a shared link
- Connect thousands of third-party applications through their APIs
- Mix manual and automated steps in one workflow and automate them gradually
- Collect input from people through a form and feed it into the automation
- Publish a workflow as a reusable template for others to clone
- Embed the automation engine in your own product through the HTTP API or an iframe
When to use Nekton / When not to
A quick filter to help you decide if Nekton is the right fit.
When to use Nekton
- Small teams and solo founders who need automation but have no developer to spare
- Operations and IT staff automating onboarding, offboarding, incident handling and hardware requests
- Support teams that want repetitive user requests handled without learning a workflow builder
- Marketers and community managers automating surveys, newsletters and social posting
- Developers who want the AI to draft the automation, with TypeScript available when it falls short
When not to use Nekton
- Anyone expecting one sentence to build a whole business: the vendor itself rules out vague, large-scale goals
- Teams whose automation must run locally or reach systems that are not exposed to the internet
- Data-heavy pipelines, since the vendor states the tool suits small to medium volumes only
- Organisations bound by GDPR paperwork, HIPAA or FISMA, none of which the vendor documents or accepts
- Buyers who need a contractual support commitment rather than a public community forum
How to use Nekton
A typical end-to-end flow, from setup to results.
- Create a free account from the homepage; no payment card is requested
- Write down the task you want handled, one step at a time, in plain English
- Let the AI turn the automatable steps into code, then refine them through the AI chat
- Follow the tool's suggestions when it proposes rewording a step it cannot automate cleanly
- Connect the third-party applications the workflow needs, through their APIs
- Add a form where the workflow has to collect information from people
- Leave the steps not worth automating as manual steps inside the same flow
- Run the workflow, put it on a schedule, or send its run link to whoever needs it
- Drop into TypeScript for any step the generated code does not cover
- Turn a finished workflow into a template to reuse internally or publish it
Pros & Cons
Pros
- Plain English is the only required input: no visual builder to learn, no developer to hire
- A permanent free plan of 300 flows a month, plus full gratuity for non-commercial and open-source projects
- Very low and fully public pricing, at 9 and 18 US dollars a month
- An enormous integration surface, with a dedicated page for each of more than 5,600 applications
- Shared run links let colleagues or customers trigger a workflow without an account
- A TypeScript escape hatch for whatever the generated automation cannot handle
- The vendor publishes what the tool does badly, which is rare and genuinely useful
Cons
- No public sign of activity since October 2023: the blog stops there and the copyright still reads 2023
- No GDPR mention anywhere, no processing agreement, no subprocessor list and no named hosting country
- Nothing is said about whether customer data trains AI models, or how to opt out of it
- The publisher is thinly identified: a Delaware mail-forwarding address, no team page, no company LinkedIn
- Support runs through a public community forum, with no contractual commitment attached
- The quota counts flow executions rather than operations, and workflows simply stop once it is exceeded
- The entire site depends on JavaScript, and API access is granted only on request
Pricing & Plans
Nekton offers a permanent free plan rather than a time-limited trial: it covers 300 flow executions per month at no cost. The cheapest paid tier, Starter, is priced at 9.00 USD per month for 1,000 flows, and the Pro tier at 18.00 USD per month for 5,000 flows. The vendor further states that use is free of charge for non-commercial and open-source projects. Upgrades may be made at any time and are charged on a pro rata basis.
- for individuals - 0 USD per month - 300 flows per month
- for small teams and projects - 9 USD per month - 1
- 000 flows per month
- for bigger teams and projects - 18 USD per month - 5
- 000 flows per month
- Free of charge for non-commercial and open-source projects
Data, GDPR & hosting
A consolidated view of how Nekton handles your data.
GDPR overview
Nekton.ai never mentions the GDPR. The term appears nowhere on the site: not in the privacy policy of 17 April 2023, not in the terms of 1 March 2023, and on no dedicated page. There is no Article 27 EU representative, no named legal basis, no enumeration of data-subject rights, no data processing agreement and no subprocessor list. This is silence rather than refusal, and the distinction matters. The policy does grant access, correction and deletion on request, through account settings or by email to info@nekton.ai, which informally covers part of the same ground. The company is registered in Delaware and the terms apply Delaware law. International transfers are explicitly anticipated, with no mechanism named to govern them. European buyers should assume nothing is documented and ask before committing regulated data.
Who owns the data?
Nekton.ai's privacy policy, last updated 17 April 2023, states that the company collects your email address, your first and last name and usage data, and that you keep the right to access, correct or delete it, either from your account settings or by writing to info@nekton.ai. The picture reverses for anything you volunteer as feedback: the terms of use, dated 1 March 2023, assign to the company all intellectual property rights in any suggestion, comment or idea you send, without acknowledgement or compensation. Neither document says who owns the workflows you build or the data your automations process, and that silence is the most significant gap in the paperwork.
Reuse rights
The terms of use grant a narrow licence: non-exclusive, non-transferable and revocable, covering only access to the Services and downloading or printing a copy of content you have properly reached, and solely for personal, non-commercial use or an internal business purpose. Everything else is withheld. Nekton's content and marks may not be copied, republished, aggregated, translated, distributed, sold, sub-licensed or exploited commercially without prior written permission. What the terms never address is the material you generate yourself: nothing is said about reusing your workflows, the code produced for you, or the results your automations return, so those outputs carry neither an explicit grant nor an explicit restriction. On the privacy side, data may be shared with service providers, group affiliates and unnamed business partners, and no subprocessor list is published anywhere.
Data retention & training
Hosting summary
Nekton names no hosting location at all. The privacy policy, dated 17 April 2023, says only that information is processed at the company's operating offices and in any other places where the parties involved in the processing are located, a formulation that identifies neither a country nor a region. It explicitly anticipates that data may be transferred to and held on computers outside your own jurisdiction, where data protection law may differ, and names no mechanism to govern that movement. No subprocessor list is published. The only jurisdictional anchor available is corporate rather than technical: the company is registered in Delaware, the privacy policy names Delaware as its country of reference, and the terms apply Delaware law. The site is served from behind Cloudflare on an anycast address, but that describes content delivery and says nothing about where data is stored. Since the product's core function passes task descriptions to OpenAI, at least one processor is implied that the documentation never acknowledges. Anyone with data-residency requirements should treat the question as undocumented and ask directly.
Things to keep in mind
Risks and trade-offs to weigh before adopting Nekton.
- The task descriptions you write are processed through OpenAI, and nothing on the site says whether that content trains models or how to prevent it
- No hosting country, region or subprocessor is ever named, so you cannot know where your data sits or who else handles it
- The privacy policy allows sharing with unnamed business partners, a clause broad enough to cover a great deal
- Automating a procedure before you understand it hides the reasoning: when a generated workflow breaks, nobody on the team may know what it was doing
- Letting AI-written code act on live systems, sending messages, moving records or posting publicly, makes an unnoticed error scale quietly
- Nothing has been publicly updated since October 2023, so building critical processes on a product whose maintenance is unconfirmed is a real operational risk
- The terms claim intellectual property over any feedback you send, and require users to be 18 while the privacy policy speaks of 13
Setup & Integrations
Technical difficulty
Low to begin with. There is nothing to install: you create an account, describe your task in plain English, and the AI produces the automation. The vendor's claim that you need neither a complex platform nor a developer holds for simple tasks, and a workflow can start out entirely manual and be automated step by step. Difficulty reappears at the edges, when authenticating and wiring up third-party APIs, writing TypeScript for steps the AI cannot generate, or integrating the HTTP API and embedded UI, which are granted on request rather than self-service.
Deployment
Integrations
Supported languages
Behind Nekton
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
How do I actually build an automation in Nekton?
What exactly counts as a flow in the pricing?
What happens if I exceed my quota?
Is there a free plan?
Do the people I share a workflow with need an account?
Which AI model is behind it?
Is Nekton GDPR compliant?
Is there an API or a mobile app?
What is the minimum age to use it?
How do I contact the publisher?
Should you pick Nekton?
Nekton makes a genuinely different bet from most automation platforms. Rather than asking you to assemble a workflow from blocks, it asks you to write down what you want in ordinary English and lets the AI produce the running code. For small teams with no developer to spare, that lowers the entry cost considerably, and the pricing reinforces it: a permanent free tier of 300 flows a month, paid plans at nine and eighteen dollars, and no charge at all for non-commercial or open-source work. The integration surface is genuinely large, the ability to hand someone a run link they can use without an account is a practical touch, and the TypeScript escape hatch keeps the ceiling reasonably high. The homepage even states plainly what the tool handles badly, which is rare enough to be worth crediting.
Two reservations deserve equal weight. The first is freshness: nothing on the site has been publicly updated since October 2023. The blog stops there, the copyright still reads 2023, and both legal documents date from early that year. The service itself still answers, payment links included, but a prospective buyer should ask directly whether the product is still being developed. The second is compliance documentation, which is close to absent. The GDPR is never mentioned, there is no processing agreement, no subprocessor list and no stated hosting country, and nothing at all is said about whether the content of your tasks, which transits through OpenAI, is used to train models. The publisher itself is thinly identified, behind a Delaware mail-forwarding address with no team page.
Nekton is worth a trial for low-stakes internal automation, where the free plan costs nothing but time. For regulated data or anything business-critical, settle the open questions first.
- Choosing a selection results in a full page refresh.
- Opens in a new window.