Transparent pricing for enterprise-grade integration

Additionally each Process or API can have at maximum 50 integration Tasks within it and a maximum of 10 integration triggers associated with it.

It is an API that is managed, secured and monitored in Frends, but the message is routed to third-party API or microservice, thus the business logic or functionality is executed elsewhere outside Frends.

Cloud deployment means that instead of deploying the integration runtimes or Frends Agents into infrastructure you manage, you will purchase them as a service from Frends where they will be hosted in Frends managed Azure infrastructure. Additionally for Cloud deployments each integration runtime or Agent will have a limitation of 250 simple integration executions per second.

An Execution is one complete run of a process or API call. Sub-processes within a main process don't count as separate executions. Each pass-through API call is counted as one execution.

Example: If your order sync process runs every 5 minutes, that's 288 executions per day (12 per hour × 24 hours).

An Agent (or Frends Agent) is the runtime component that executes your integrations. Agents can be:

• Cloud Agents: Hosted and managed by Frends
• Self-Service Agents: Deployed in your own infrastructure
• HA Agent: Cloud or Self-Service Agent in high-availability (HA) and load-balancing mode. Available in Enterprise Tier.

Example: You might have one agent in your on-premises data centre for sensitive data processing and cloud agents for SaaS integrations.

High-Availability (HA) Agent, available in Enterprise Tier, is a Frends Agent offering redundancy, load-balancing and is able to execute more operations per second than a basic Agent. Technically this means that Frends has two Agents running exactly the same APIs or Processes and if one of the pair is down, the other Agent still performs the joint duties. As both are active all the time and they share the load, HA Agent offers also more performance than a standard Agent.

Technically HA Agent can include more than two standard Agents inside, but those setups are based on quote on Frends Cloud Agents or customers can deploy them as Self-Serve Agents where there is no limitation other than maximum of 100 Agents per one Frends Tenant. See more here.

They mean Environments built within Frends to enable seamless deployments throughout the pipeline from Development to Production. For fully customized pipelines you can define as many environments as you'd like.

Environment is a part of the pipeline and contains integration runtimes or Frends Agents which actually execute the integrations deployed to said Environment. Note that for Hybrid Deployments you can install as many Agents within a single Environment on your own infrastructure as you'd like to enable Hybrid Scenarios.

High-Availability support is the ability to deploy multiple integration runtimes to share the same load and failover in case of a disaster scenario. Note High-Availability support for the Frends Agents is only available on the Enterprise tier.

Frends utilizes Azure Resources to provide the service and those services are configured to be isolated logically from one-another. With fully dedicated premium resources utilized Azure Resources are upgraded to a premium version in Azure and are fully dedicated for each Customer.

The limitation means the total number of API and Integration executions allowed within a month across all Environments and integration runtimes or Agents for a given Frends Service. Note that the limitation is also dependent on the integration runtimes capability to scale enough to reach the given limit. The limitation is also not a hard-limitation meaning that if you go over the limitation Frends will not block exceeding executions, but a move to a higher Tier is instead highly recommended.

Frends Credits are tokens consumed when using Frends-hosted AI models for intelligent automation tasks.

Conversion: 1 Frends Credit ≈ 3000–6000 AI tokens (input, output and caching combined) depending on the chosen AI model. E.g. GPT 4o consumes less tokens than GPT5.

How are Credits consumed?

Credits can be consumed partially based on the complexity of the AI task:

• Simple email categorization: ~0.12 credits
• Complex document analysis: ~2–5 credits
• Large-scale text processing: ~10+ credits

Credit consumption is calculated at the end of the process execution, so e.g. small amounts of AI tokens consumed inside a Frends loop are not calculated as Frends Credits until the process end. This prevents the problem rounding up might make small loops expensive.

One Frends Credit costs €0.01 and the consumption-based invoicing occurs per 1,000 Frends Credits (€10 per 1,000 credits).

Credit Management

• See the credit consumption of single execution during the development of a Frends Process
• Monitor consumption via the admin portal
• Set monthly consumption limits to control costs
• Option to use your own AI models instead of Frends-hosted ones

Example: If you process 1,000 support emails daily using AI categorization (0.12 credits each), you'd consume approximately 3,600 credits per month.

With the normal Frends Tiers you are not allowed to re-sell any part of the Service or Solutions and you should instead look at Frends ISV Partner tier.

The number of Processes in use per month is calculated on the last day of every month by looking at unique Frends Processes, APIs and Pass-through APIs in use outside the Development Agent Group in the Development Environment. The Development Agent Group is defined as the first Agent Group in the Environment view of Frends.

For the sake of clarity both parties agree that the Development Agent Group is also identified in the User's browser under the path /Environment/Default/Default. This means that if you have an integration process that is deployed to Development, Test and Production at the same time that is still only counted as a single Process for invoicing. You can see the number of Processes yourself directly from the Frends platform at any time.

Yes. You can on a month-by-month basis move up or down in Tiers provided you are not using features from a Higher tier that are not available in a lower tier.

Yes, the number of processes in use is measured continuously every month and can decrease if you delete existing integrations or APIs.