Question 1

What is Avintra Community Edition?

Accepted Answer

A source-available edition of Avintra that you deploy on your own infrastructure. It includes all core platform features in a single workspace, with SSO enablement, Docker-based deployment, BYOK and community support.

Question 2

What can I use Community Edition for without paying?

Accepted Answer

What can I use Community Edition for without paying?

Avintra may be used free of charge. Without a Commercial License you may use, copy and modify it for:

Internal use within your organisation, including your affiliates
Consulting, implementation, integration, customisation, training and support services for clients, including for a fee
Personal use, learning and experimentation
Academic research, teaching and non-profit use

Hosting it for third parties, or selling it or a product that includes it, needs a Commercial License.

Question 3

How is the Enterprise plan different?

Accepted Answer

Enterprise includes everything in Community, plus multiple workspaces, full audit logs and governance, dev, staging and prod environments and dedicated support with SLA. Enterprise features are licensed under the separate Avintra Enterprise License.

Question 4

Does Community Edition use credits?

Accepted Answer

No. Credits only apply to the Starter and Pro plans on Avintra Cloud. Community Edition runs on your own infrastructure with your own model keys (BYOK).

Question 5

Which license is Community Edition released under?

Accepted Answer

The Avintra Community License: a modified version of the Apache License 2.0 with additional conditions covering commercial use, logo and copyright information, Enterprise features and contributions. Where those conditions conflict with the Apache License 2.0, the conditions prevail. All other rights and restrictions follow the Apache License 2.0 (apache.org/licenses/LICENSE-2.0).

Question 6

Can my group companies use it under the same terms?

Accepted Answer

Yes. Internal use covers your organisation and its affiliates: entities that control, are controlled by, or are under common control with you, where control means more than 50% ownership or the power to direct management.

Question 7

Can I charge clients for services built on Avintra?

Accepted Answer

Yes. Consulting, implementation, integration, customisation, training and support services for clients are permitted without a Commercial License, including when you charge for them. What needs a Commercial License is hosting Avintra for those clients, or selling it as part of your product. See the next question.

Question 8

When do I need a Commercial License?

Accepted Answer

You need a Commercial License from Aventisia, a separate written agreement, to offer Avintra or a derivative work to third parties as a hosted, managed or software-as-a-service offering, whether single-tenant or multi-tenant. You also need one to sell, license or distribute Avintra or a derivative work for a fee, or to include it in a product or service that is sold or licensed to third parties.

Question 9

Can I modify the source code?

Accepted Answer

Yes. You may use, copy and modify Avintra for any of the permitted uses. A derivative work is treated like Avintra itself: offering it to third parties as a hosted service, or selling it, needs a Commercial License. Two parts stay off-limits: the logo and copyright information in the frontend, and the Enterprise Files and their license key checks.

Question 10

Can I resell Avintra or build a competing platform on it?

Accepted Answer

No. No license, including a Commercial License, permits you to resell, sublicense or rent Avintra itself, or to offer Avintra or a derivative work, under any name, as a standalone or competing AI, agent, workflow automation or application-building platform, unless Aventisia expressly agrees otherwise in writing.

Question 11

Can I remove the Avintra logo or white-label the product?

Accepted Answer

Not under the Community License. When using the Avintra frontend you may not remove or modify the logo or copyright information in the Avintra console or applications, unless a Commercial License expressly permits white-labelling. The frontend means all components in the web/ directory of the source code, or the "web" image when running Avintra with Docker.

Question 12

What are Enterprise features, and how are they licensed?

Accepted Answer

Source files with ".ee." in their filename, every file inside a directory whose name contains ".ee", and any compiled or built form of those files are Enterprise Files. They are not licensed under the Community License, only under the Avintra Enterprise License (LICENSE_EE.md). You may not remove, disable, bypass or modify any license key check that controls access to Enterprise features, wherever it sits in the code, or distribute code or instructions intended to do so. Doing so terminates your rights under the Community License.

Question 13

Can I contribute to Avintra?

Accepted Answer

Yes. Contributions are made under the Avintra Contributor License Agreement (CLA.md), in place of section 5 of the Apache License 2.0. Aventisia may use contributions for commercial purposes, including in its cloud and Enterprise offerings.

Question 14

Can the license terms change?

Accepted Answer

Aventisia may change the license terms of future versions of Avintra. Versions already released keep the license they were released with.

Question 15

How do I get a Commercial License?

Accepted Answer

Contact us at contact@aventisia.com and we will take you through the options.

Question 16

What are credits?

Accepted Answer

Credits are the consumption units used across the Avintra platform. Every time an agent is executed, a small number of credits are consumed based on the resources used — so you only pay for actual execution, never for idle licenses.

Question 17

How do I purchase credits?

Accepted Answer

Starter and Pro plans include a monthly or annual credits allowance. You can top up more credits directly inside Avintra application, or contact our sales team for offline upload.

Question 18

How are credits calculated?

Accepted Answer

Credits are calculated based on two factors: Agent Execution Time — Credits are consumed based on the duration for which your AI agent or workflow runs. Platform-Managed AI & Services — If your agent uses Aventisia-managed services, such as LLMs or other third-party APIs provided through Aventisia, additional credits are consumed to cover the usage of those services. If your agent uses your own API keys for AI models or third-party services, credits are charged only for the agent's execution time, with no additional consumption for AI services. In simple terms: Runtime only — When using your own AI/API keys. Runtime + AI service usage — When using Aventisia-provided AI models or third-party services. This transparent model ensures you only pay for the compute and services your automation actually consumes.

Question 19

Do credits expire?

Accepted Answer

Plan-allocated credits refresh with your billing cycle. Top-up credits remain valid for 12 months from purchase.

Question 20

How many credits do I need for my use case?

Accepted Answer

It depends on complexity of your Agent and APIs/Models consumed. Benchmark your agent with a few sample runs — most teams converge on a steady-state rate within the first few runs.

Question 21

Can I top up credits anytime?

Accepted Answer

Yes, credits reflect instantly once the payment is made through the Avintra application.

Question 22

What happens if I run out of credits?

Accepted Answer

Running agents pause if there are no credits left. Top up or upgrade to resume execution — no data or workflow state is lost.

Question 23

Are credits transferable between tenants?

Accepted Answer

No, credits cannot be transferred between tenants.

Question 24

Is there a minimum purchase requirement?

Accepted Answer

Credits can only be purchased in multiple of 500, so 500 can be considered as minimum purchase requirement.

Question 25

Can I get a refund on unused credits?

Accepted Answer

No, unused credits are non-refundable.

Question 26

How do I track credit usage?

Accepted Answer

The Tenant Admin has access to Usage dashboards which provide detailed break-down of credits consumed by Agents, Users, Models and Services — with exportable reports.

Question 27

Do credits apply to Enterprise plans?

Accepted Answer

No. Credits are only applicable to the Starter or Pro plans, where Avintra is consumed from Aventisia's cloud and customers can leverage Aventisia-managed AI models, LLMs, and other third-party services without bringing their own credentials. For Enterprise deployments, the credit model does not apply. These deployments are governed by the commercial terms of your agreement, such as platform subscription, execution-hour commitments, or other contracted usage metrics.

Question 28

What are Execution Hours?

Accepted Answer

Execution Hours represent the actual time during which an AI agent or workflow is actively executing within the Avintra platform. Execution time starts when an agent or workflow begins execution and ends when the execution completes, regardless of whether it is triggered manually, through an API, by a schedule, an event, or a chat interaction.

Question 29

How is execution time measured?

Accepted Answer

For every execution, Avintra records: Execution Start Time and Execution End Time. The difference between these timestamps is counted towards your consumed Execution Hours.

Question 30

Is execution time measured to the second?

Accepted Answer

Yes. Execution Hours are measured with second-level precision and aggregated across all executions during your billing period.

Question 31

Does agent build and testing consume Execution Hours?

Accepted Answer

Yes. Execution Hours include all workflow executions, including: Testing while building an agent, Debug executions, Test runs, Trial executions before deployment, Production executions

Question 32

Are chat agents charged differently?

Accepted Answer

No. For conversational AI agents, execution is measured for each user interaction/exchange. The time taken for processing that interaction/exchange contributes to your Execution Hours.

Question 33

Does waiting for an LLM response count?

Accepted Answer

Yes. Execution time includes the complete duration from the start of execution until the final response is returned. This includes time spent: Executing workflow logic, Calling AI models, Retrieving knowledge, Calling APIs, Waiting for external systems to respond, Processing outputs.

Question 34

If the workflow calls multiple AI models, is all time included?

Accepted Answer

Yes. Execution Hours include the total elapsed execution time irrespective of how many AI models, APIs, databases, or integrations are invoked during the execution.

Question 35

If multiple agents are executed within one workflow, how is execution calculated?

Accepted Answer

Execution Hours are calculated based on the actual execution time of each agent. If agents execute sequentially, their execution time naturally contributes to the total workflow execution. If agents execute in parallel, the execution time consumed by each running agent is counted independently.

Question 36

Are failed executions counted?

Accepted Answer

Yes. Execution time consumed before an execution fails or is terminated is included.

Question 37

Are cancelled executions counted?

Accepted Answer

Yes. Any processing completed before cancellation contributes to Execution Hours.

Question 38

Are idle workflows charged?

Accepted Answer

No. Only active execution time is counted. There is no charge for workflows or agents that are deployed but idle.

Question 39

Is there a limit on the number of agents?

Accepted Answer

No. Enterprise customers can create an unlimited number of: AI Agents, Reusable Blocks, Knowledge Bases, Skills, Connectors. Execution Hours are the primary consumption metric.

Question 40

Is there a limit on the number of users?

Accepted Answer

No. Enterprise licensing includes unlimited users. There is no per-user licensing fee.

Question 41

Can multiple users use the same agent simultaneously?

Accepted Answer

Yes. There is no restriction on concurrent users. Each execution contributes independently to total Execution Hours.

Question 42

What happens to unused Execution Hours at the end of the billing cycle?

Accepted Answer

Unused Execution Hours do not roll over to the next billing cycle and will expire at the end of the current billing period. To help customers optimize costs and avoid purchasing more capacity than required, Avintra recommends selecting a subscription tier that closely matches your expected monthly usage. If your execution requirements increase during the billing period, you can simply purchase additional Execution Hours at the same per-hour rate applicable to your subscribed volume tier. This approach provides two key benefits: Cost optimization - Avoid paying upfront for execution capacity that may go unused. Flexibility - Scale up seamlessly as your AI adoption grows without committing to a higher volume tier. Example if you are on Enterprise Plan of 30,000 Execution Hours/year and your Actual Usage is 32,350 Execution Hours. You can purchase the additional 2,350 Execution Hours at the same per-hour rate as your 30,000-hour subscription tier. This model ensures that customers only pay for the capacity they actually use while retaining the flexibility to scale their AI workloads as needed.

Question 43

Why does Avintra use Execution Hours instead of user or bot licensing?

Accepted Answer

Traditional AI and automation platforms often charge based on the number of users, bots, or named licenses, which can result in paying for unused capacity. Avintra's execution-based model aligns cost with actual usage. You pay only when your AI agents are actively performing work, making it easier to scale AI across the enterprise without incurring additional per-user or per-agent licensing costs and completely eliminating idle capacity.

Feature	Avintra Cloud		Self-hosted
Feature	Starter$0/mo	Pro$48/mo	CommunityFree	EnterpriseCustomMost popular
Agents
Intelligent automation agents
Chat agents
Voice agents	Coming soon	Coming soon	Coming soon	Coming soon
Generative agents
Core features
All integrations (nodes)
Knowledge base vectorization
MCP tool calling
Skills calling
Agents & prompts library
Create custom Spaces
Human & agent queues
Model setup & token consumption
Agent Risk Registry
Save / use templates
Workflow templates
Agent-to-Agent (A2A) calling
Save / call blocks (reusable components)
Connect LLMs using Aventisia key
Connect your on-prem LLMs
Bring your own key (BYOK)
Call custom code (JS / Python)
External storage integration
Triggers: webhooks, queues, email, chat
Environments (dev / staging / prod)
Version control for agents
Archive agents
Automatic retries
Top-up credits			Not applicable	Not applicable
Debugging
Execution logging
Debug in editor
Live execution tracking
Desktop assistant
Error handling
Audit logs
Security
Encrypted secrets store
Key store
SSO SAML and LDAP
Enforce 2FA across the instance
Collaboration
Users	3	Unlimited	Unlimited	Unlimited
Share agents
View users live on canvas
View live updates by other users
Broadcast notifications in your tenant
Communicate via comments
Insights
Insights & history
Credit usage			Not applicable	Not applicable
BYOK consumption analytics
Execution hours tracking
Space/Agent/User/Date filtering
History / logs (max days)	1 day	30 days	1 day	365 days / custom
Support
In-app service tickets
In-app change requests
Quick Steps to get started upon first login
In-app guides
Email support
Dedicated support with SLA
Billing & contract
Pay / upgrade online
Pay by invoice
Custom contract
Sales-assisted procurement

Just pay for execution.

Two ways to consume Avintra.

SaaS platform

Platform in your environment

Choose your plan.

Starter

Pro

Community

Enterprise

A full feature comparison across every plan.

Free to use. Clear on the limits.

Free under the Community License

Needs a Commercial License

Not permitted under any license

Licensing, credits and execution hours.

Build smarter. Innovate faster.