Diaflow
Diaflow is an AI-native workflow automation platform for teams and companies. It combines autonomous AI agents, a visual flow builder, multi-model chat, vector search, documents, tables and storage in one governed workspace billed through monthly credits.
What is Diaflow?
Diaflow is an AI-native workflow automation platform published by Diaflow Pte. Ltd., a Singapore company founded in 2024 by Jonathan Viet Pham, Lai Pham and Anh Doan. Its positioning is agentic rather than rule-based: instead of rigid trigger-action chains, it deploys AI agents that decide and act. The enterprise page sums it up bluntly: "Diaflow doesn't just automate tasks - it thinks, delegates, and completes work." Several agents can be coordinated by an orchestrator, for example a research agent, a writing agent and an editing agent working on the same deliverable.
The product is a suite of six building blocks sharing one workspace. Flows is the visual builder, where an automation can be assembled in three ways: drag-and-drop with no code, a plain-English description that assembles the nodes for you, or a Python node for custom logic. Chat gives a whole team access to several models, including GPT-5, Claude Opus, Gemini Pro, Llama and Mistral, through providers such as OpenAI, Anthropic, Google, Cohere, xAI, DeepSeek, Replicate, OpenRouter, Straico, Azure and Ollama, with an Intelligent mode that routes each request according to complexity, cost, speed and quality. Vectors turns files and links into a searchable knowledge base for retrieval-augmented generation, Drive centralizes storage, Pages generates and rewrites documents, and Tables is a no-code relational database that forms, APIs, webhooks and agents can fill. A private AI LLM range keeps inference inside Diaflow's own private cloud, with no data shared with third parties.
Connectivity is central: 200+ native apps are claimed, 156 integration pages are listed in the sitemap, and a universal API connector covers the rest. Governance sits on top, with RBAC, SAML 2.0 and OpenID Connect SSO, MFA, audit logs and a monitoring dashboard, on a compliance base of SOC 2 Type II, HIPAA and GDPR backed by a public trust portal. Everything runs on monthly credits: each flow execution, agent interaction and API call consumes them. Diaflow claims up to 80% savings on AI costs and more than 10,000 active users. Deployment is SaaS on AWS, or on-premises and private cloud for enterprise customers.
What it does
- Build AI workflows on a visual canvas, without writing code
- Deploy autonomous AI agents that plan, execute, check and iterate on a task
- Query several LLMs from a single company-wide chat interface
- Connect more than 200 third-party apps, or any other API through the HTTP Request node
- Turn files and links into a vector knowledge base for retrieval-augmented answers
- Generate documents, slides, images, audio and video from a prompt
- Read handwritten text and extract data from images and documents with OCR and vision models
When to use Diaflow / When not to
A quick filter to help you decide if Diaflow is the right fit.
When to use Diaflow
- Operations and business teams that want to replace manual, multi-step processes with automated flows without writing code
- Companies that would rather pool LLM access in one governed workspace than pay for scattered individual AI subscriptions
- Developers and technical builders who need a Python node, an HTTP Request node and API keys inside the same canvas
- Regulated organizations that require SOC 2 Type II, HIPAA and GDPR commitments plus on-premises or private-cloud deployment
- Knowledge and support teams building a retrieval-augmented knowledge base from their own files and links with Diaflow Vectors
When not to use Diaflow
- Mobile-first users: no iOS or Android app is published, despite an Apple App Store EULA addendum in the compliance center
- Teams that need a localized product: the interface and the documentation are English only, and support runs mainly in English and Vietnamese
- Occasional or seasonal users: monthly credits do not roll over and any unused balance expires at the end of the cycle without compensation
- Buyers who expect a money-back guarantee: no refund is issued when the customer cancels the subscription
- Solo users who need more than the free Basic limits of 1 seat, 5 flows, 1 GB of storage and a single LLM per flow
How to use Diaflow
A typical end-to-end flow, from setup to results.
- Create an account on the Diaflow platform: the free Basic plan requires no credit card
- Open your workspace and choose how to build: drag-and-drop canvas, natural-language prompt, or Python node for custom logic
- Describe the automation in plain English, for example "When a new lead arrives from HubSpot, enrich the data, score it, and notify the sales team on Slack"
- Pick a trigger for the flow: webhook, form, schedule (cron) or HTTP call
- Add the apps and models the flow needs, and fall back on the HTTP Request node for anything not natively supported
- Feed Diaflow Vectors by dropping files or links into a root or group folder, then select that source to build the knowledge base
- Generate API keys from the workspace API keys page, notably for the Webhook node and external calls
- Run the flow, then monitor usage and buy extra credits from the dashboard with the Add Credits button when needed
- Get help from the AI assistant in the circle at the bottom right of the site, the GitBook documentation or the Discord community
- For an enterprise rollout, follow the four-step path: Discovery and Solution Fit, Demos and Customization, Team Training and Rollout, then Support and Optimization
Pros & Cons
Pros
- One subscription covers automation, AI chat, a vector database, documents, tables and storage instead of several separate tools
- Three build levels in the same canvas: no-code drag-and-drop, natural language, and Python for custom logic
- Documented compliance base, with SOC 2 Type II, HIPAA and GDPR claims and a public trust portal
- Sovereignty options that are rare at this price: Diaflow private cloud, on-premises deployment and bring-your-own API keys
- A permanent free plan with no credit card required, offering 500 credits per month and 5 flows
- Very large integration catalog that can be checked page by page: 156 integration pages are listed in the sitemap for a claimed 200+ apps
- Granular pricing with 15 credit tiers, from 4,000 to 1,150,000 credits per month, and a documented training opt-out with a stated address and processing time
Cons
- Monthly credits do not accumulate or roll over: any unused balance expires at the end of the billing period, with no compensation and no refund
- No refund when the customer cancels; a pro-rata refund only applies if Diaflow terminates the contract without cause
- Diaflow reserves the right to change the credits allocated to free Basic accounts at any time and without prior notice
- The subprocessor list is maintained but never published, being available on request only, and the SOC 2 Type II report is accessible under NDA only
- The Article 27 EU representative is mentioned twice but never named, and the minimum age contradicts itself: 13 in the terms, 16 in the privacy policy
- Outdated content on some pages: the Vectors FAQ still advertises a paid plan from USD 199 per month against USD 16 on the pricing page, and invokes the EU-US Privacy Shield, invalid since 2020; the about page og:image still points to http://localhost:3000
- No mobile app is published despite an Apple App Store EULA addendum, and some enterprise security guarantees are marked "Only valid for on-premises or private cloud customers"
Pricing & Plans
Diaflow offers a permanent free plan. The Basic tier is billed at USD 0, requires no credit card and includes 500 credits per month. Paid entry starts at USD 16 per month with the Pro Starter plan on annual billing, that is USD 192 per year, or USD 20 per month on monthly billing; the annual commitment is advertised with a 20% discount. Pro is USD 32 per month on annual billing, USD 40 monthly, and Business is USD 320 per month on annual billing, USD 400 monthly. Enterprise pricing is available on quotation through the demo form or sales@diaflow.io. Additional credits can be purchased at any time from the dashboard, and exceeding a quota suspends workflows until the next cycle unless the plan is upgraded or extra capacity is bought. Three billing points deserve attention: unused credits expire at the end of the cycle, cancellation by the customer does not trigger a refund, and a free trial converts into automatic billing at its term without prior notice.
- 500 credits per month
- 100 flow runs
- 1 user
- 5 flows
- 2
- 000 records per table
- 1 GB of storage
- a single LLM per flow
- 4
- 000 credits per month
- 800 flow runs
- 100 flows
- flow builder
- 10 GB of file storage
- built-in vector database
- secure agentic chat and access to advanced AI models
- 8
- 000 credits per month
- 800 flow runs
- same feature set as Pro Starter
- 80
- 000 credits per month
- everything in Pro plus unlimited flows
- 50 GB of storage
- 10 GB of vector database
- connection of your own AI models
- team and role management
- governance and monitoring dashboard
- on-premises deployment
- advanced security options
- dedicated compute resources
- deep customization
- tailor-made enterprise applications and use of your own API key
- the pricing page exposes 15 credit steps
- from 4
- 000 to 1
- 150
- 000 credits per month
- with prices climbing from USD 16 to USD 4
- 000 per month on annual billing (USD 20 to USD 5
- 000 on monthly billing)
Data, GDPR & hosting
A consolidated view of how Diaflow handles your data.
GDPR overview
Diaflow explicitly claims to be fully compliant with the General Data Protection Regulation. The privacy policy covers access, rectification, erasure, restriction, portability, objection, withdrawal of consent and the right to complain to a supervisory authority; requests go to support@diaflow.io and are answered within the legal deadline, typically 30 days. Transfers outside the EEA rely on Standard Contractual Clauses with a UK Addendum. A Data Processing Agreement is available to all customers on request from legal@diaflow.io, and customers are notified of a breach within 72 hours, in line with Article 33. Two gaps remain: the Data Protection Officer is presented as the GDPR contact but is never named, and the EU representative is mentioned twice without ever being identified. CCPA/CPRA rights and the Singapore PDPA are also acknowledged.
Who owns the data?
Diaflow positions itself as a Data Processor acting on behalf of its customers, who remain the Data Controllers, so your content stays yours. The publisher states that it does not sell personal data and that customer data used inside AI agent workflows is logically isolated and never used to train models for other customers. In practice, you can export your workflows in JSON format while the subscription is active, and you can ask for your Workspace data to be returned within 30 days after the contract ends. Diaflow's own rights over your content are limited to running the service, subject to the training rules described in the terms.
Reuse rights
Diaflow processes data to deliver the service, run AI tasks, improve the product, communicate with users, market with consent, protect against fraud, meet legal obligations and meter credit consumption. Requests can be routed to third-party model providers such as OpenAI, Anthropic and Google, while the Chat page states that conversation data is never sent to those providers for training. The terms allow Diaflow to use anonymized and aggregated data to improve, train, fine-tune or benchmark AI models unless you opt out in writing to privacy@diaflow.io, a request handled within 15 business days. Identifiable customer content is not used for training without prior explicit consent, and Google API data is excluded from model training and from advertising under the Limited Use policy. AWS is named as the infrastructure subprocessor.
Data retention & training
Hosting summary
Diaflow's cloud service runs on a multi-tenant architecture hosted by Amazon Web Services, with VPC isolation, network segmentation and multi-availability-zone deployments. The terms state that the services are operated from Singapore, and that data may be transferred, stored and processed in Singapore or in any other jurisdiction where Diaflow and its subprocessors operate. Data is encrypted at rest with AES-256 and in transit with TLS 1.2 or above, with keys managed by AWS KMS and rotated automatically. Backups run daily across several geographic zones and are kept for 7 days. The database can be hosted by Diaflow, be the customer's own database, or be a third-party database that Diaflow reaches as a proxy with server-side credentials. Enterprise customers can switch to an on-premises deployment or a private cloud on AWS, GCP or Azure to meet data-residency requirements. One reserve: no specific AWS region is published, so the exact country of the datacenters is not guaranteed by the site.
Things to keep in mind
Risks and trade-offs to weigh before adopting Diaflow.
- Credit economics work against irregular use: unused monthly credits expire without compensation, and Diaflow can change the allocation of free Basic accounts at any time and without prior notice
- No refund is granted when you cancel, and a free trial converts into automatic billing at its term without prior notification, so the end date needs to be diarized
- Transparency gaps you cannot audit yourself: the subprocessor list is available on request only, and the SOC 2 Type II report is shared under NDA only
- The Article 27 EU representative is never named and the minimum age contradicts itself between the terms (13) and the privacy policy (16), which weakens the compliance story for European buyers
- Some legal and pricing content is outdated: the Vectors FAQ still quotes USD 199 per month and relies on the EU-US Privacy Shield, invalid since 2020
- Data may leave your country: the services are operated from Singapore, and part of the enterprise security guarantees are valid only for on-premises or private cloud customers
- Autonomous agents act on live systems through hundreds of connectors: flows deployed without review can propagate an error at scale, and teams that stop checking agent output gradually lose the reflex of verifying it
Setup & Integrations
Technical difficulty
Low to moderate. Signing up takes minutes and the Basic plan needs no credit card. Flows are built by drag-and-drop or by describing them in plain English, and a template library organized by department gets a non-technical user to a first automation without writing a single line of code. Technical depth is optional: Python nodes, HTTP Request nodes, API keys and webhooks are there when needed. Enterprise rollouts are heavier, with SAML 2.0 or OpenID Connect SSO to configure and a guided four-step onboarding with a customer success manager. GitBook documentation and a Discord community support both paths.
Deployment
Integrations
Supported languages
Behind Diaflow
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Does Diaflow have a free plan?
Do unused credits roll over to the next month?
What happens if I exceed my credit quota?
Can I change or cancel my plan at any time?
Are there discounts or special offers?
Can I connect an application that is not in the catalog?
Can I use Diaflow with my own documents and data?
Is my chat data sent to OpenAI, Anthropic or Google for training?
Can Diaflow be deployed on-premises?
Which certifications does Diaflow hold?
Should you pick Diaflow?
Diaflow is an integrated bet. Rather than selling one more automation tool, it packs AI agents, a visual builder, multi-model chat, a vector database, documents, tables and storage into a single team subscription. That makes it far more compelling for an organization trying to consolidate scattered AI subscriptions than for an individual chasing a single feature. The compliance foundation is unusually solid for a young company: SOC 2 Type II, HIPAA and GDPR claims, a public trust portal, a DPA available to every customer on request, and a training opt-out with a stated address and a 15-business-day commitment.
The reservations are contractual rather than technical. Monthly credits never roll over and expire without compensation, cancellation triggers no refund, the subprocessor list stays available on request only, the SOC 2 report is NDA-only, and the Article 27 EU representative is invoked but never named. A few pages have aged badly: the Vectors FAQ still quotes USD 199 per month against USD 16 on the pricing page and relies on the Privacy Shield, invalid since 2020, while the minimum age contradicts itself between the terms and the privacy policy.
The entry ticket stays low, with a permanent free plan without a credit card and then USD 16 per month on annual billing, so a serious evaluation costs little. Read the credit mechanics before committing, because they, not the headline price, drive the real bill. Diaflow Pte. Ltd. was founded in 2024 and announced an unquantified seed round in 2025: promising and well documented, but still young. Teams with a data-residency requirement should look straight at the on-premises or private-cloud option rather than at the shared SaaS.
- Choosing a selection results in a full page refresh.
- Opens in a new window.