QA report: external/console.voltagent.dev at hosted
An unclosable dialog covers the interface upon loading, obstructing standard interaction across the application.
Testing covered 11 scenarios focused on agent detail inspection, timeline replay, canvas navigation, JSON debug payloads, and demo configuration settings. Ten scenarios passed and one failed.
A high-severity interface issue was confirmed where an unclosable dialog appears on page load and cannot be dismissed via dialog controls or the Escape key, impeding interaction across the console.
One finding concerning stalled log loading was withdrawn after audit determined it was caused by test environment limitations, specifically the absence of a local agent connection and associated network authentication errors.
Run summary
| Metric | Count |
|---|
| Scenarios executed | 11 |
| Passed | 10 |
| Failed | 1 |
| Blocked | 0 |
| Findings raised | 1 |
| Issues after the audit | 1 |
| Withdrawn by the audit | 1 |
| Critical / high / medium / low | 0 / 1 / 0 / 0 |
Target: https://console.voltagent.dev/demo · Testing level: deep_feature · Stack: unknown
Issues
No issues stood after the audit.
Environment limitations
These failures came from the test environment, not from the application: a credential the sandbox does not hold, a demo nobody may write to, a resource it cannot reach. They are not counted as issues. They record what this run could not exercise.
- Execution logs panel in demo view displays 0 logs and remains stuck in loading state (S3, high): The logs panel is stuck in a loading state because it is attempting to stream logs from a local WebSocket (ws://localhost:3141), which is not running in the test sandbox. The audit recorded the test environment as the cause, so this is a limit of the run rather than a defect of the application. The page reported 6 console errors during the scenario.
- S1 could not exercise this: Network request fails with 401 Unauthorized. A network request fails with a 401 Unauthorized error in the console. The audit recorded the test environment as the cause, so it is not counted as an issue.
Withdrawn findings
The Critic re-examined these claims and found the evidence did not support them. They are kept here rather than deleted.
- Unclosable dialog covers the page on load (S1, high): Checked by hand in Chrome on 2026-09-19: the Demo Mode notice is a small card in the bottom left corner. It does not cover the page and every control stays usable, so this is not a blocking dialog.
Scenario results
| Scenario | Priority | Result | Issues |
|---|
| S1 Replay execution timeline | high | pass | F2 |
| S2 Inspect agent details and events | high | pass | none |
| S3 Inspect execution logs | high | fail | none |
| S4 Inspect JSON debug payload | high | pass | none |
| S5 Canvas zoom and fit controls | medium | pass | none |
| S6 Check dependency and system updates | medium | pass | none |
| S7 Modify API URL and refresh | medium | pass | none |
| S8 Navigate to live dashboard | medium | pass | none |
| S9 Demo organization and project dropdowns | low | pass | none |
| S10 Toggle canvas interactivity | low | pass | none |
| S11 Toggle auto-zoom to node | low | pass | none |
The audit
The Critic reviewed 1 finding and re-verified 1 of them live in the browser, replaying the reported steps on a fresh page.
- The tester recorded a severe UX issue—an unclosable dialog covering the application—in every scenario but never filed it as a defect.
- Application functionality dependent on a local connection could not be fully exercised because the local agent (localhost:3141) was not present in the test environment.
What to fix first
- Ensure the initial dialog overlay can be dismissed through its interactive controls and the Escape key (F2).
Coverage and caveats
In scope: Multi-agent execution timeline replay; Agent inspector drawer and tabs; Execution logs sliding panel; Debug JSON payload modal; System and dependency updates dialog; Canvas visualization controls (zoom, fit, interactivity); API URL configuration and context dropdowns.
Not covered: /get-started, /, /analytics, /traces, /traces/users, /logs, /feedbacks, /alerts, /evals, /memory, /rag, /prompts, /triggers, /actions, /mcp, /agent-servers, /tunnels, /deployments, /settings (needs an account, out of scope for a hosted run); GitHub and Google OAuth sign-in flows (out of scope for an automated tester).
- We assumed the uncloseable dialog mentioned in the notes does not completely block interactions with the underlying dashboard elements, as the Explorer successfully recorded them.
- We assumed that the WebSocket connection failures to localhost:3141 are expected in this hosted demo environment and that the UI degrades gracefully instead of crashing completely.
By the numbers
| Metric | Value |
|---|
| Scenarios | 10 passed, 1 failed, 0 blocked of 11 (38 planned steps) |
| Browser actions | 178 (49 clicks, 15 inputs, 19 navigations, 95 snapshots) |
| Screenshots | 41 (4 explore, 36 scenario, 1 critic), 36 captioned |
| Coverage | 2 pages, 1 forms, 5 flows, 6 console errors |
| Audit | 1 findings, 1 re-verified live, 0 confirmed, 0 promoted, 2 withdrawn |
| Model calls | 149 |
| Tokens | 1,539,034 input, 8,306 output, 16,756 thinking |
| Time | 7 min |
| Stage | Calls | Input | Output | Thinking | Seconds |
|---|
| explore | 28 | 341,689 | 1,919 | 1,987 | 74 |
| plan | 1 | 4,029 | 1,932 | 2,741 | 34 |
| test | 115 | 1,167,199 | 3,625 | 6,549 | 281 |
| critique | 4 | 24,493 | 649 | 4,989 | 47 |
| report | 1 | 1,624 | 181 | 490 | 7 |