Drafter AI
Latvian no-code platform for building AI tools and automations by combining 100+ data sources and ML models, aimed at teams without ML engineers. The product application itself was unreachable at the time of collection, on 11 August 2026.
What is Drafter AI?
Drafter AI is a no-code platform for building AI applications and automations, published by Phonal Technologies SIA in Riga, Latvia. Its home page sums the promise up as Launch Your AI App in Minutes. Zero Code. Instead of writing code, the user assembles building blocks, data sources and machine learning models, into workflows that then run automatically. The site claims an inventory of more than 100 data sources and ML models, and advertises a 90% reduction in development costs.
The AI blocks documented on the site cover a wide range: text generation; text analysis for key phrases, entities, sentiment and topics; custom classification; machine translation; image recognition; speech recognition for interviews, calls and meetings; PDF recognition; web data extraction; public search through Google or Bing; internal search; structuring of raw data; and interpretation of raw numbers. An internal knowledge base can be fed with PDFs, videos or web content, indexed and then queried by the AI, while prompt templates are reusable and prompt sequences can be automated.
Output takes three forms. A branded web application can be published on the customer's own domain, with a dedicated portal on the Scale tier. Every workflow can also be called through the developer API documented on the GitBook site, where API calls are generated automatically for each workflow. Finally the platform positions itself as a step inside existing automation chains, through Make.com, Zapier, Airtable and Google Sheets. Local deployment is offered for sensitive documents.
The same engine is marketed in four commercial variations: the general platform on the home page, Drafter AI for Sales, an internal knowledge system, and a white-label GPT chatbot. A service activity sits alongside it, building GenAI proofs of concept, with 48 PoCs claimed and fourteen days quoted as the average.
Several figures come from the vendor and cannot be verified independently: 1,000+ sales teams, 16,000+ personalised emails generated per day, 100+ client companies and 50+ companies said to have launched AI tools. The site also carries the disclaimer ChatGPT is a trademark of OpenAI. Drafter AI is not affiliated with OpenAI.
What it does
- Build an AI application without writing any code
- Chain several AI tasks into a single workflow that runs automatically
- Connect internal documents as a knowledge base the AI can draw on
- Process hundreds of inputs in a single batch run
- Expose any workflow through a developer API
- Publish the application under your own brand and domain
When to use Drafter AI / When not to
A quick filter to help you decide if Drafter AI is the right fit.
When to use Drafter AI
- Marketing, sales and operations teams that want to assemble AI automations without ML engineers on staff
- No-code agencies looking to add AI building blocks to the deliverables they sell to clients
- Product teams that want to test an AI feature before committing engineering time to building it
- Sales operations handling CRM enrichment and personalised outreach at volume, the use case the vendor documents most
- Support and knowledge teams building an internal knowledge base or a white-label GPT chatbot on top of their own documents
When not to use Drafter AI
- Anyone who needs access right now: the product application was unreachable when the tool was checked on 11 August 2026
- Organisations that require a signed DPA, a formal subprocessor list or a SOC 2 or ISO certification, none of which is published
- Teams whose processing must stay inside the European Union: the privacy policy announces transfer to and processing in the United States
- Buyers who need a contractual guarantee that their data will not be used to train models, a point on which the site says nothing either way
- Mobile-first users, since no iOS or Android application exists, and anyone under 18, excluded by both the terms and the privacy policy
How to use Drafter AI
A typical end-to-end flow, from setup to results.
- Create an account and verify the email address, which the documentation asks to be a professional one
- Wait for an instance to be deployed automatically, then open the workspace
- Start from the template workflows made available at first launch
- Assemble a workflow from the available building blocks: Inputs, Actions and Flow Controls
- Pick among the claimed 100+ data sources and ML models to compose each step
- Save reusable prompt templates and chain prompt sequences where the task requires it
- Upload PDFs, videos or web content to build the internal knowledge base the AI will draw on
- Run the workflow in batch, up to 200 entries in two minutes according to the sales page, and work with the results as tables
- Publish a branded web app: apply your branding, connect a custom domain and invite your team
- Go further through the developer API or the Make.com integration; note that the sign-up button points to app.drafter.ai, unreachable at the time of collection
Pros & Cons
Pros
- Broad functional coverage in a single tool: text, image, speech, PDF, web extraction, public search and internal search
- Several output modes: a branded web app, a developer API, or delivery inside Make.com and Google Sheets
- A single credit system whatever the building block used, described and quantified on the pricing page
- Public, itemised entry pricing, which is uncommon in this segment
- A 30-day free trial written into the terms of service, not merely used as a sales argument
- An identifiable company registered in Latvia, with a published postal address and telephone number
- White-label and optional local deployment for sensitive contexts, backed by structured public documentation with credit costs given per task type
Cons
- The product application was unreachable at the time of collection: app.drafter.ai returned HTTP 530, a Cloudflare 1016 origin DNS error, on three attempts on 11 August 2026, so the advertised free trial could not be reached; blog.drafter.ai did not respond either
- Maintenance signals have been stopped for a long time: status page not updated since 17 March 2023, last changelog post on 26 May 2023, documentation showing 'Last updated 3 years ago', most recent sitemap timestamp 17 December 2024
- Legal documents are dated and out of step with the product: terms of 27 August 2021 still describe a 'cloud-based advertising texts generating platform', and the privacy policy dates from 9 May 2023
- Pricing contradicts itself between pages: USD 24, 79 and 1,999 per month on the pricing page against USD 199 per month on the knowledge and chatbot pages, for different scopes
- Integration claims are inconsistent: Slack and Zapier are presented as available on the commercial pages but marked 'soon' in the documentation
- No DPA, no formal subprocessor list and no certification such as SOC 2 or ISO is published; nothing is stated about model training on customer data or any opt-out; and transfer to and processing in the United States is announced without naming a transfer mechanism
- A single contact channel, the obfuscated address support@drafter.ai, and no mobile application on either store
Pricing & Plans
There is no permanent free plan. A 30-day free trial is granted, once per account, under article 8 of the terms of service. The lowest published entry point is USD 24.00 per month for the Use tier, which covers 1,000 operations per month, the default AI apps, the Slack and Make.com integrations and a single user. The Customize tier starts at USD 79.00 per month for 500,000 AI credits, up to 5 users, unlimited apps and multiple roles. The Scale tier is offered from USD 1,999.00 per month against a struck-through USD 10,000.00, and includes 1,000,000 AI credits, unlimited users, a white-label portal, custom API integrations, developer API access, dedicated consulting time and optional local deployment. Extra credits are sold at USD 79.00 per 500,000. Billing combines a fixed monthly component with a pay-per-use component on credits consumed, invoiced by the number of 500,000-credit packages started during the month. Two other pages, Knowledge and GPT Chatbot, advertise a different entry point, USD 199.00 per month for 1,000,000 credits, with a Scale tier on quotation. Payment is processed through Stripe, no pro-rata refund is granted on early termination, and a 3% penalty applies for each month of late payment.
- Use — USD 24 per month
- Customize — from USD 79 per month
- Scale — from USD 1
- 999 per month
- displayed against a struck-through USD 10
- 000
- Basic
- on the Knowledge and GPT Chatbot pages — USD 199 per month for 1
- 000
- 000 credits
- Scale
- on the Knowledge and GPT Chatbot pages — on quotation
- unlimited credits
Data, GDPR & hosting
A consolidated view of how Drafter AI handles your data.
GDPR overview
Phonal Technologies SIA is established in the European Union, in Riga, Latvia, and its terms apply Latvian law. The privacy policy, last updated on 9 May 2023, carries a dedicated section titled 'Your Data Protection Rights Under General Data Protection Regulation (GDPR)' listing access, update, erasure, rectification, objection, restriction, portability and withdrawal of consent, together with the right to complain to a data protection authority in the European Economic Area. Rights are exercised through a single address, support@drafter.ai. Two nuances matter. The site never writes that it is 'GDPR compliant': it enumerates the rights and undertakes to allow them. And no data processing agreement is published or offered, no data protection officer is named, and no Article 27 representative is designated, which is expected for an EU-established company. CalOPPA and CCPA sections are also present.
Who owns the data?
Under article 15 of the terms of service, dated 27 August 2021, the customer owns and keeps all rights, title and interest, including intellectual property rights, in Customer's Data. Beyond what the terms expressly license, no right, title or interest passes to Phonal Technologies. The publisher receives a non-exclusive, royalty-free, transferable licence to use and publish the content, solely to deliver the service and for the customer's benefit, and may not resell, rent, assign or share customer data without prior written agreement, except to perform its own obligations. Phonal keeps ownership of its technology, its methods and its platform integration credentials. The customer remains solely responsible for the content generated and warrants holding the necessary rights.
Reuse rights
The terms leave the customer free to reuse its own data and the results produced: it retains all rights over Customer's Data and needs no permission from Phonal Technologies to exploit them, in exchange for being solely responsible for what is generated and for warranting that it holds the rights to the source material. On the publisher's side, the privacy policy of 9 May 2023 presents Phonal as data controller and lists the personal data collected: email address, first and last name, cookies and usage data, the latter covering IP address, browser type and version, pages visited, date and duration of visits, device identifiers and diagnostic data. The stated purposes are providing and maintaining the service, notifying changes, enabling interactive features, providing support, analysing and improving the service, monitoring usage, detecting and fixing incidents, performing contractual obligations and billing, sending account and subscription notices, and sending offers and news, which can be declined. Session, preference, security and advertising cookies are used, with remarketing declared through Google Ads, Bing Ads, Twitter and Facebook, analytics through Google Analytics, CI/CD through GitHub and payments through Stripe, the publisher storing no card data itself. The policy states: We do not sell your personal information for monetary consideration. Do Not Track signals are declared honoured and a CCPA opt-out is offered. One silence deserves to be named: nowhere does the site say whether customer data is used to train models, neither to confirm it nor to rule it out.
Data retention & training
Hosting summary
The privacy policy states that data may be transferred to and held on computers located outside the user's jurisdiction, and says so explicitly: If you are located outside United States and choose to provide information to us, please note that we transfer the data, including Personal Data, to United States and process it there. The publisher gives a general undertaking to take reasonable steps to ensure secure processing, but names no transfer mechanism: no standard contractual clauses, no Data Privacy Framework. In parallel, the pricing page announces that customers can choose their preferred storage geography and their third-party data providers, and local deployment is mentioned on the Scale tier and on the sales page for sensitive documents. These two statements are never reconciled. No hosting country or region other than the United States is named, no infrastructure subprocessor is identified and no hosting certification is claimed. The publisher itself is established in Riga, Latvia.
Things to keep in mind
Risks and trade-offs to weigh before adopting Drafter AI.
- The application behind the 'Try for Free 30 days' button, app.drafter.ai, returned HTTP 530 and a Cloudflare 1016 origin DNS error on three attempts on 11 August 2026, and blog.drafter.ai was unreachable on two attempts: the buying path advertised on the site leads nowhere at present
- The status page drafterai.statuspage.io displays 'All Systems Operational', but its own last update dates from 17 March 2023: that display says nothing about the current state of the service
- Other maintenance signals have stopped as well: last changelog post on 26 May 2023, GitBook documentation showing 'Last updated 3 years ago', most recent sitemap timestamp on 17 December 2024
- The terms of service, dated 27 August 2021, still describe the service as an advertising-copy generation platform, a far narrower scope than what is sold today: the contract you would sign does not match the product presented
- Prices contradict each other from one page to another, USD 24, 79 and 1,999 per month on the pricing page against USD 199 per month on the knowledge and chatbot pages, and Slack and Zapier are shown as available commercially but marked 'soon' in the documentation
- Traction figures are unverifiable, with 1,000+ teams, 16,000+ emails per day, 100+ companies and 48 proofs of concept claimed, and a scarcity notice reading '3 deals left' is displayed on the Scale tier
- No security certification is claimed despite the 'Perfect Security' argument used on the sales page, a single obfuscated contact address (support@drafter.ai) is published, and robots.txt blocks indexing of the terms and privacy pages
Setup & Integrations
Technical difficulty
Low on paper. The platform is positioned as no-code and states that no ML engineer is required: sign up, verify a professional email address, wait for an instance to be deployed automatically, then start from template workflows. Workflows are built from Inputs, Actions and Flow Controls blocks. Going further calls for more skill: developer API, Make.com integration, custom domain connection, with dedicated consulting time included on the Scale tier for complex cases. One practical obstacle overrides all of this: no instance can be created, since app.drafter.ai was unreachable at the time of collection.
Deployment
Integrations
Behind Drafter AI
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Drafter AI.
Frequently asked questions
Is Drafter AI accessible today?
What exactly is a credit?
How does the pay-per-use billing work?
Is there a free plan?
Do you need to know how to code?
Can you choose where the data is stored?
Which integrations are actually available?
Is there a mobile application?
What is the minimum age to use the service?
Should you pick Drafter AI?
On paper Drafter AI is a substantial product: a registered Latvian company, two complete legal pages, a postal address and a telephone number, three itemised pricing tiers, public GitBook documentation, a demonstration video and named case studies, with broad functional coverage and an entry price that is low for this segment. What the collection observed on 11 August 2026 points the other way: the product application at app.drafter.ai returned an HTTP 530 Cloudflare error on three attempts, the blog did not respond, the status page has not been updated since March 2023, the changelog stops in May 2023 and the most recent sitemap timestamp is December 2024. The legal framework is dated and incomplete as well, with terms of service from 2021 that still describe a much narrower product, no DPA, no formal subprocessor list and complete silence on whether customer data is used to train models. The reasonable course is therefore to check the accessibility of the service and the current state of the publisher directly with support@drafter.ai before committing to anything.
- Choosing a selection results in a full page refresh.
- Opens in a new window.