Flowise

Choose Flowise if you built your AI system using Flowise. Mibo integrates directly with Flowise and can automatically collect trace data, so there’s minimal setup required.

Active and passive testing

Flowise supports both active and passive testing:

• Active testing works with Agentflow v2 only. Mibo needs the detailed execution data that Agentflow v2 provides. If you’re using an older flow type, consider migrating to Agentflow v2.
• Passive testing works with any Flowise flow type (chatflows, agentflows v1 or v2). As long as your flow sends trace data to Mibo, it can be evaluated. See the Flowise Trace Format for how to send traces.

Setting up your connection

1. Add your agentflow

   Go to your project, click Add Platform, and select Flowise.

2. Enter your Chatflow URL

   Paste the full URL of your Flowise flow. Mibo automatically extracts the flow ID from the URL, so you don’t need to find it separately.

3. Set up authentication (if needed)

   If your Flowise instance requires an API key, enter it here. If your instance is open (no authentication), you can skip this step.

4. Choose a message template

   The message template defines the request body Mibo sends to Flowise. It uses placeholders that get replaced with real values each time a test runs:

   Placeholder	Replaced with
   {input}	The test case’s input text
   {VARIABLE_NAME}	The value of VARIABLE_NAME from the test case’s context
   {external_id}	(Optional) The trace ID for this execution

   Mibo provides two presets optimized for different flow types:

   • Chatflow: for standard Flowise chatflows. Sends { "question": "{input}" }.
   • Agentflow: for Flowise Agentflow v2. Includes the additional fields that agentflows expect.

   Pick the one that matches your setup. You can also customize the template if your flow expects a different format.

   Example: if your test case has a context variable like SESSION_USER, you can pass it along:

   {
     "question": "{input}",
     "overrideConfig": {
       "sessionId": "{SESSION_USER}"
     }
   }

   With a test case context of { "SESSION_USER": "usr-123" }, Mibo sends "sessionId": "usr-123" on every request for that test.

   See context variables to learn how to define variables in your test cases.

   Tip

   The {input} placeholder is replaced with the input_utterance from each test case when Mibo runs your tests. If a test case uses custom_body instead, the entire template is replaced. See The Test Architect: Input types for details on how test inputs work with your message template.

5. Test your connection

   Click Test Connection to verify Mibo can reach your Flowise instance and get a response.

How Mibo works with Flowise

When running active tests, Mibo sends inputs through Flowise’s prediction API, the same way your users interact with the flow. The response is automatically extracted from the correct field, so you don’t need to configure a response path manually.

Trace data and the Failure Matrix

One of the biggest advantages of using Flowise with Mibo is automatic trace collection. When you use Agentflow v2, Flowise provides detailed execution data for every interaction, including:

• Which nodes executed and in what order
• What tools the agent called and with what inputs
• How the agent reasoned through each step
• The final response and how it was assembled

Mibo uses this data to power the Failure Matrix, which breaks down issues into four stages: Router, Arguments, Tool, and Response. This helps you pinpoint exactly where something went wrong (not just that it went wrong).

Tip

Since Flowise Agentflow v2 includes trace data automatically, you’ll get full access to the Failure Matrix and stage-level analysis out of the box. No extra configuration needed.

Sessions

Mibo supports Flowise session IDs. If your flow uses sessions to maintain conversation history, Mibo will create a unique session for each test execution. This ensures tests don’t interfere with each other or with your real users’ conversations.

Which Flowise flow type should I use?

Flow Type	Active Testing	Passive Testing	Trace Detail
Agentflow v2	Yes	Yes	Full: nodes, tools, reasoning
Agentflow v1	No	Yes	Varies
Chatflow	No	Yes	Basic: input and output only

If you’re starting fresh or can choose your flow type, Agentflow v2 gives you the best experience with Mibo. You’ll get active testing, automatic traces, and the full Failure Matrix.

If you’re already running a chatflow or older agentflow, you can still use Mibo through passive testing. Just make sure your flow sends traces.

Using the Test Architect with Flowise

When creating tests with the Test Architect, you can upload your Flowise chatflow JSON export to help the AI understand your agentflow’s structure. The AI will automatically detect testable components (agents, tools, chains) and suggest test targets.

To export your flow: in Flowise, open your agentflow, click the settings icon, and export as JSON. Then attach the downloaded file in the Test Architect chat, or paste its contents directly.

Note

Only JSON files are supported (not screenshots, PDFs, or other document formats). The file size limit is 500 KB.

Troubleshooting

• “Connection failed”: make sure your Flowise instance is running and accessible from the internet. If it’s running locally, it won’t be reachable by Mibo.
• “Authentication error”: check that your API key matches what’s configured in your Flowise instance settings.
• “Flow not found”: verify the URL is correct and the flow is deployed (not just saved as a draft).
• No trace data appearing: make sure you’re using Agentflow v2 for active testing. Older flow types don’t provide the execution data Mibo needs.
• Unexpected responses: check that you selected the right message template (Chatflow vs. Agentflow). Using the wrong template can cause your agentflow to misinterpret the input.