Aerogram
Aerogram is an enterprise AI orchestration platform that turns expert workflows and SOPs into governed AI apps any employee can run from a form. It combines a no-code visual builder, an internal app portal and multi-model access.
What is Aerogram?
Aerogram is a web-based collaborative workspace for building AI workflows, published by Cleverse Corporation Co., Ltd., a company established in Bangkok. The product has two faces that work together. The first is a visual builder, the Think-board Canvas, where cards are dropped onto a board, connected to one another and configured model by model. The second is an enterprise AI portal: once a workflow is finished it is published as an AI App, and colleagues who never touch the canvas pick it from an internal catalogue, fill in a form and click Run App. That is the thesis the home page states plainly, No coding. No Prompting, and it answers three problems it names openly: an uneven skills gap between employees, inconsistent results from improvised prompting, and AI usage that management cannot see.
Under the hood the platform is model-agnostic. The documentation listed one hundred models in July 2026: 44 for text and reasoning, 26 for image and 30 for video. They span GPT, Claude, Gemini, DeepSeek, Llama, Qwen, GLM, Grok, Perplexity Sonar and Kimi on the text side, and Flux, Ideogram, Recraft, Stable Diffusion, SeeDream, Luma, Kling, Pika, PixVerse, Veo, Sora, Wan and Hunyuan for image and video. Around them sit reusable building blocks: Modules, an agentic Knowledge Base, File Storage, a Slide Generator, Published Content, Chat and App in Chat, a Scanner Card and an Audio & Transcriber.
Governance is the second pillar. SSO, RBAC, separated environments, member and department management, human checkpoints, activity logs and an attributed revision history are part of the offer, alongside the ISO/IEC 27001:2022 certification displayed on the home page. Consumption is metered in credits whose cost varies with model complexity and output size, and organisations are structured into business units, User and Creator licences and seat tiers per model family, under a Team or an Enterprise plan.
The home page illustrates the idea with departmental apps: Resume Pre-Screening for HR, a Competitive Analysis SWOT, a CRM Insights Digest for sales, an Employee Skill Gap Analysis, a Market Positioning Map Generator and a branding architecture app for marketing. The publisher also proposes a five-step engagement, Discover, Prioritize, Build, Launch & Measure and Scale, rather than leaving customers to build alone. Access is browser-based, mobile included, with the same multi-model capabilities.
What it does
- Build multi-step AI workflows by drag and drop, without writing code
- Publish a workflow as an AI App that business users run from a simple form
- Chain models together, wiring the output of one card into the input of the next
- Generate text, images and video from a catalogue of around one hundred models
- Query an internal knowledge base (RAG) to ground answers in company documents
- Pull data from third-party apps such as Gmail, Slack, Notion, Google Sheets, Jira and HubSpot into a board or a chat
- Trace every run through activity logs and a timestamped, attributed revision history
When to use Aerogram / When not to
A quick filter to help you decide if Aerogram is the right fit.
When to use Aerogram
- Corporate operations, finance, legal, HR and support teams looking to automate repetitive, procedure-driven work
- Employees with no prompt-engineering skills, who need to run AI through a form rather than through a chat box
- IT, risk and compliance leaders who need RBAC, SSO, activity logs and revision history to govern how AI is used
- Agencies and creators producing text, image and video content who want many models behind a single interface
- Educators and training teams packaging repeatable AI exercises and course material for their students
When not to use Aerogram
- Buyers who need a published price before engaging: no amount, tier or currency appears anywhere, every plan goes through a sales contact
- Developers looking for a public API: none is documented and no API reference is available
- Teams working outside English and Thai, the only two languages of the interface and the documentation
- Users who need a native mobile app: there is no iOS or Android application, mobile access goes through the browser
- Individuals, solo freelancers and anyone under 20: the terms set a minimum age of 20 and the product is built around organisational portals, licences, business units and shared credits
How to use Aerogram
A typical end-to-end flow, from setup to results.
- Create an account through Google or Microsoft OAuth (terms §2.1)
- Accept the privacy policy and the terms of service in the consent dialog shown on first access
- Open a new board in the visual canvas, or start from a public template in the prompt-template catalogue
- Add input cards: text, document, presentation, scanner or Knowledge (RAG)
- Choose the AI model for each card and adjust its parameters
- Connect the cards so that the output of one becomes the input of the next
- Add a third-party connector from a card or from the chat with Add Connector, then confirm the authorisation
- Publish the finished workflow as an AI App, so business users only fill in a form and click Run App
- Invite collaborators, set their permission level, and revoke access when it is no longer needed
- Track consumption through the credit balance shown at the bottom left of the application home page
Pros & Cons
Pros
- Around one hundred models behind a single interface and a single bill, with no provider lock-in
- End users never write a prompt: they fill in a form and click Run App
- Governance built in: RBAC, SSO, activity logs, revision history and human checkpoints
- ISO/IEC 27001:2022 certification displayed by the publisher
- Zero-Data-Retention required from every model provider, and no self-hosted model, so customer content trains no proprietary model
- Expertise is capitalised: reusable workflows, modules, public templates and a shared knowledge base
- A five-step engagement is offered, from Discover to Scale, so customers are not left to build on their own
Cons
- No public pricing at all: not a single amount, tier or currency on the website or in the documentation
- No public API and no API documentation
- Interface and documentation available in English and Thai only
- No native iOS or Android application
- No DPA published, no sub-processor list and no hosting country or region disclosed
- Entirely Thai legal framework (PDPA), with no mention of the GDPR and no EU representative
- Credit-based consumption whose cost varies with the model and the output size, making spend hard to anticipate
Pricing & Plans
No price is published. Neither the website nor the documentation states an amount, a tier or a currency, and no permanently free plan is described; the documentation indicates that both the Team plan and the Enterprise plan are obtained by contacting Aerogram. A free trial is advertised through the Start Free Trial buttons on the public platform pages, and usage is then metered through a credit system. Under section 9.1 of the terms, billing is monthly or annual depending on the cycle chosen at purchase, with pro rata charging on an upgrade and effect at the next cycle on a downgrade. Section 9.3 states that amounts paid are generally non-refundable, and section 9.2 reserves the publisher's right to change its rates with reasonable notice.
- Team plan
- price available on request from Aerogram
- all Team plan features plus business units and credit management per business unit
- price available on request
- User license and Creator license
- the two licence types distinguished in the documentation
- Seat tiers per model family (text
- image
- video)
- which can be set department by department
- Free trial
- advertised through the Start Free Trial buttons on the public site
Data, GDPR & hosting
A consolidated view of how Aerogram handles your data.
GDPR overview
There is no mention of the GDPR anywhere in Aerogram's privacy policy or terms of service: the regulation is never named, no EU representative under Article 27 is appointed, no data protection officer is identified and no DPA is published. The framework actually declared is Thailand's Personal Data Protection Act B.E. 2562 (PDPA). Section 5 of the privacy policy lists the rights it grants: withdrawal of consent, access and copy, portability, rectification, erasure, restriction, objection and complaint, with the Personal Data Protection Committee in Thailand named as the supervisory authority. Section 14 of the terms makes Thai law applicable and gives Thai courts exclusive jurisdiction. On security, the home page displays an ISO/IEC 27001:2022 certification, with AES-256 or equivalent encryption at rest and TLS 1.2 or above in transit.
Who owns the data?
Under section 3.1 of the terms, you keep ownership of everything you create, upload or submit, prompts and files included, and you own AI outputs to the extent they are legally ownable, with no assurance that generated material qualifies for copyright protection. Section 3.2 grants the publisher only a limited, non-exclusive licence to access, store, process and transmit your content, and solely as far as necessary to deliver the service. Section 8.2 states that Aerogram claims no intellectual property in customer content, while the service, its software and its trademarks remain the property of Cleverse Corporation Co., Ltd. or its licensors. Suggestions and feedback you send may be reused without compensation or attribution (§8.3).
Reuse rights
End users can reuse what they generate without asking Aerogram for permission: no commercial restriction is placed on outputs, and section 8.2 states that the publisher claims no intellectual property over customer content. Several reservations apply. Section 3.3 makes the user solely responsible for the content submitted and for holding the rights necessary to use it. Section 4.4 warns that AI outputs may be inaccurate or incomplete and leaves review entirely to the user, while section 4.5 specifies that nothing generated by the Service constitutes legal, financial, medical, or any other form of professional advice. Sharing stays under the user's control (§6): invited collaborators may view, edit, comment on, copy or export content according to the permissions granted, and access can be revoked at any time. Any advertising or marketing use of personal data requires an active opt-in (privacy policy §3.2).
Data retention & training
Hosting summary
Aerogram does not say where its data is hosted. Neither the privacy policy nor the terms of service names a hosting country or region, and no list of sub-processors is published. What is stated is the jurisdiction: the publisher, Cleverse Corporation Co., Ltd., is established in Bangkok, and section 14 of the terms makes Thai law applicable with exclusive jurisdiction to the courts of Thailand. On the technical side, stored customer content is encrypted at rest with AES-256 or an equivalent (privacy policy §2.2), and traffic is encrypted in transit with TLS 1.2 or above (§3.3). Aerogram runs no self-hosted model: inference is delegated to native providers, directly or through third-party gateways cited as examples, OpenRouter and Vercel AI Gateway, with customer content streamed rather than stored thanks to the Zero-Data-Retention commitment required of those providers. The domain resolves through Cloudflare (188.114.96.0, anycast), which indicates the delivery network and not the storage location. In practice, the residency question cannot be answered from public documents and has to be raised directly with the publisher.
Things to keep in mind
Risks and trade-offs to weigh before adopting Aerogram.
- AI outputs may be inaccurate, incomplete or inappropriate, and section 4.4 of the terms leaves the review entirely to the user
- Nothing generated counts as legal, financial or medical advice (§4.5), yet the polished app format can make an output look authoritative
- Skill erosion: employees who never write a prompt gradually lose the ability to judge what the model returns
- Frozen judgement: a published workflow encodes one expert's reasoning and can quietly become obsolete if nobody reviews it
- Third-party connectors give the platform access to Gmail, Drive, Slack or Notion data according to the permissions granted
- Shared content stops being private: collaborators can copy or export it depending on their permissions (§6)
- Hosting locations and sub-processors are undocumented, so there is no way to know where data resides, and the publisher commits to neither availability nor fitness for purpose (§10)
Setup & Integrations
Technical difficulty
Low for the end user, moderate for the administrator. Signing up goes through Google or Microsoft OAuth, with no password to manage; workflows are built by drag and drop without code, and public templates are ready to use. A business user only fills in a form and clicks Run App, with no prompting skill required. Third-party connectors need an OAuth authorisation per application, handled from the Connectors page. Administration is heavier: members, business units, credits, User and Creator licences and seat tiers. The publisher also offers to design and deploy the workflows itself, in five steps, with no self-build required.
Deployment
Integrations
Supported languages
Behind Aerogram
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What is Aerogram?
Do I need to know how to write prompts?
How many AI models are available?
Is the pricing public?
Is there a free trial?
Is my content used to train AI models?
How long is my data kept?
Which data protection framework applies?
Is there a minimum age?
Is there a mobile application?
Should you pick Aerogram?
Aerogram does not try to be another model; it positions itself as the orchestration and governance layer above them. Its real proposition is organisational: an expert's method is captured once in a visual workflow, published as an AI App, then run by colleagues who fill in a form instead of improvising a prompt. Everything that follows from that, reusable modules, an internal catalogue, activity logs, an attributed revision history, RBAC and SSO, is what makes the promise credible for a company that has to account for its AI usage. Unified access to roughly one hundred text, image and video models metered in credits under a single bill, the ISO/IEC 27001:2022 certification and the Zero-Data-Retention commitment demanded of every model provider all point the same way.
The reservations are just as clear. Pricing is entirely opaque: not one amount, tier or currency is published, so any evaluation has to start with a sales conversation. There is no public API, and the interface and documentation exist only in English and Thai. The legal framework is Thai from end to end, PDPA, Thai law, Thai courts, with no GDPR provision, no Article 27 representative, no published DPA and no disclosed hosting location, which European buyers will need to settle contractually before anything else. The product is also young: the domain was registered in April 2025 and the public site is still uneven, with older pages coexisting alongside the current home page.
The natural audience is medium to large organisations wanting to industrialise documented procedures with AI while keeping a visible, auditable trail of what is run and by whom. Smaller teams, developers looking for an API and buyers who need published prices will find the fit harder.
- Choosing a selection results in a full page refresh.
- Opens in a new window.