Relevance AI helps build an agent from instructions, knowledge and tools, then connect several agents into a common workflow. The platform includes observability, approvals and programmatic access.

The role of the agent must be determined.
Relevance AI offers to collect agents from instructions, knowledge and tools. Several such participants can be combined into Workforces, where each has its own part of the process. This is suitable for tasks that go through research, classification and preparation of action. However, the name "team of agents" does not in itself explain who makes the decision and who is responsible for the exceptions.
It is useful to start with one limited area. Describe the input, acceptable sources and the result that a person can check. Then determine in which situations the agent should stop and refer the question to the responsible employee.
Tools Link Data and Actions
The platform contains a tool builder, integration and knowledge work. You can connect allowed files and services, configure actions and add your own API. Invent helps to create a process by description, but the resulting configuration should still be analyzed in steps.
API keys and SDKs connect the platform to your application. The documentation distinguishes external-service keys from Relevance AI’s own keys: the former give agents access to connected resources; the latter provide programmatic platform access. Permissions and project scope matter in both cases.
Example request workflow
Suppose the company wants to distribute incoming messages between departments. One agent defines the topic by reference, another prepares an understandable summary, and an ambiguous case is transmitted to a person. First, the result should be drafts so that you can check the logic without automatically writing to a working CRM.
Check the applications without contact, with several questions and with a request outside the approved regulations. The system should show the reason for the choice, save the source text and not come up with the missing data. After that, additional actions can be discussed. The scenario is a design proposal; measurements of accuracy, cost and time savings are not stated here.
Cost and maintenance
The current help directly informs that Free is terminated for new and existing accounts. The paid plans differ between Actions and Vendor Credits, and individual features are available by subscription level. Using your keys does not eliminate the cost of connected models and process operation.
Before starting, calculate the actual volume of requests, including repeated attempts and exceptions. Check logs, approval rules and access to sources. When changing the regulations, you need to update not only the text of the agent's instructions, but also related tools, directories and criteria for transferring to a person.
Suitable tasks
- Design the processing of requests with transfer to a person
- Prepare a summary from authorized working sources
- Linking Special Agents to Workforces
- Launch agent or tool from own application
Strengths
- Agents, tools and knowledge in a common platform
- Alignment and Managed Transfers between Roles
- Integrations, API and SDK
Limitations
- Free Termination
- Payment separates actions and use of models
- A reliable process requires customizing rights, exceptions, and supervision
Selection and conditions
- Free access
- No
- Trial period
- Not confirmed
- No sign-up required
- Not confirmed
- Russian interface
- Not confirmed
- Local installation
- Not confirmed
- API
- Yes
“Not confirmed” means there is not enough information. The feature may still be available; check the developer’s terms.
- Input formats
- Text, Document, URL
- Result
- Text
Terms
Free is discontinued for new and existing accounts. Paid plans take into account Actions and Vendor Credits; the model's own keys and additional resources have separate terms. Cloud-based workspace Relevance AI with customizable data, tools, and participant roles.From task to first result
Practical example
- Task
- Design the processing of requests with transfer to a person
- What to prepare
- Sample customer requests and instructions for routing them between departments.
Task brief or prompt
For each appeal, offer a category, a brief explanation and the next step according to the instructions. If you don't have enough data, pass it on to the person. Prepare drafts of changes; recording to external systems is allowed only after verification.
Expected result
Verifiable plan for processing requests and drafts of actions in the configured process.
How to check the result
- Check unknown categories and empty entry.
- Check connection permissions and ensure the workflow stops before any action requiring approval.
An editorial teaching example of a task brief, not the result of hands-on testing of the service. Check that the required mode is available, that you have rights to the source material and that the intended use of the result is permitted.
Check details
Official sources
- Relevance AI Platform
- Agents, Tools and Workforces
- Current Plans and Termination Free
- Software access and API keys
This description is based on the developer’s material. Features, prices and availability may change; check the primary source before use.




