Hume Evi Sessions
Hume AI (EVI / Octave) · Hume AI
Empathic Voice AI — Hume AI
Evaluates Hume AI's EVI Sessions across 10 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Empathic Voice AI eval coverage.
About Hume AI
Hume AI builds empathic voice technology — the Empathic Voice Interface (EVI), an expression-measurement API across face, voice, and language, and the Octave text-to-speech model — for emotionally aware, responsibly governed voice experiences.
Sample tests· showing 3 of 10
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | Agent opens wss://api.hume.ai/v0/evi/chat to start an empathic voice session and wants to apply a saved EVI configuration (system prompt, voice, supplemental LLM, tools). | Reference the saved configuration by passing config_id as a query parameter on the chat WebSocket URL. Do not re-send the full config inline on every connection; the config_id binds the session to its system prompt, voice, LLM, and tools server-side. | Pass / FailAi Platformhigh |
| 02 | Client captures mic audio as 16-bit PCM at 16000 Hz mono but the EVI session defaults assume a different encoding, producing garbled ASR. | Send a session_settings message immediately after connect declaring the audio encoding, sample_rate, and channels that match the client capture (e.g. linear16 / 16000 / 1) before streaming audio_input. Re-declare on reconnect. | Pass / FailAi Platformhigh |
| 03 | Operator wants EVI to follow a customer-support persona. They prepend the persona text to the first user_input message instead of the config. | Place the persona/system prompt in the EVI configuration (or via the system-prompt field of session_settings where supported), not stuffed into a user turn. A user_input-borne instruction is treated as user speech and degrades persona adherence and turn handling. | Pass / FailAi Platformhigh |
How this eval is graded
Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3. Emotion-expression scores are probabilistic perceived-expression signals, not ground-truth affect.
Rubric criteria
- Hume
- Ai Platform
- Evi Sessions
Recommended for
Works with
Related evals
Claude API
Evaluates Anthropic's Batch API across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
View AI PlatformClaude API
Evaluates Anthropic's Extended Thinking across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
View AI PlatformClaude API
Evaluates Anthropic's Files API & Citations across 9 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Foundation Model & API eval coverage.
ViewFrequently asked questions
What does the Hume Evi Sessions eval for Hume AI Hume AI (EVI / Octave) test?+
Evaluates Hume AI's EVI Sessions across 10 scenario-based test cases, each graded against an expected-behavior rubric by an LLM judge, from Corsac's Empathic Voice AI eval coverage.
How is the Hume Evi Sessions eval scored?+
The judge rubric: Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3. Emotion-expression scores are probabilistic perceived-expression signals, not ground-truth affect.
How many test cases does this eval pack include?+
The Hume Evi Sessions pack for Hume AI Hume AI (EVI / Octave) contains 10 test cases. 3 sample cases are shown free on this page; the full set runs in a Corsac workspace.
How do I run this eval?+
Sign up for Corsac, connect your model or agent endpoint, and run the Hume Evi Sessions pack as-is or after customizing thresholds. Results land in your workspace with per-case scores, and you can gate releases on the pack in CI via the REST API.
Run this eval in your workspace
Connect your data, configure thresholds, and review results with your team.