29
29
30
30
EVI uses this text to generate spoken audio using our proprietary
31
31
expressive text-to-speech model. Our model adds appropriate emotional
32
- inflections and tones to the text based on the user’ s expressions and
32
+ inflections and tones to the text based on the user' s expressions and
33
33
the context of the conversation. The synthesized audio is streamed
34
34
back to the user as an [Assistant
35
35
Message](/reference/empathic-voice-interface-evi/chat/chat#receive.AssistantMessage.type).
75
75
Base64 encoded audio input to insert into the conversation.
76
76
77
77
78
- The content of an Audio Input message is treated as the user’ s speech
78
+ The content of an Audio Input message is treated as the user' s speech
79
79
to EVI and must be streamed continuously. Pre-recorded audio files are
80
80
not supported.
81
81
@@ -166,7 +166,7 @@ types:
166
166
Once this message is sent, EVI will not respond until a [Resume
167
167
Assistant
168
168
message](/reference/empathic-voice-interface-evi/chat/chat#send.ResumeAssistantMessage.type)
169
- is sent. When paused, EVI won’ t respond, but transcriptions of your
169
+ is sent. When paused, EVI won' t respond, but transcriptions of your
170
170
audio inputs will still be recorded.
171
171
custom_session_id :
172
172
type : optional<string>
@@ -250,7 +250,7 @@ types:
250
250
system_prompt :
251
251
type : optional<string>
252
252
docs : >-
253
- Instructions used to shape EVI’ s behavior, responses, and style for
253
+ Instructions used to shape EVI' s behavior, responses, and style for
254
254
the session.
255
255
256
256
@@ -301,7 +301,7 @@ types:
301
301
Third party API key for the supplemental language model.
302
302
303
303
304
- When provided, EVI will use this key instead of Hume’ s API key for the
304
+ When provided, EVI will use this key instead of Hume' s API key for the
305
305
supplemental LLM. This allows you to bypass rate limits and utilize
306
306
your own API key as needed.
307
307
tools :
@@ -367,7 +367,7 @@ types:
367
367
Parameters of the tool. Is a stringified JSON schema.
368
368
369
369
370
- These parameters define the inputs needed for the tool’ s execution,
370
+ These parameters define the inputs needed for the tool' s execution,
371
371
including the expected data type and description for each input field.
372
372
Structured as a JSON schema, this format ensures the tool receives
373
373
data in the expected format.
@@ -521,7 +521,7 @@ types:
521
521
type : string
522
522
docs : >-
523
523
User text to insert into the conversation. Text sent through a User
524
- Input message is treated as the user’ s speech to EVI. EVI processes
524
+ Input message is treated as the user' s speech to EVI. EVI processes
525
525
this input and provides a corresponding response.
526
526
527
527
@@ -540,7 +540,7 @@ types:
540
540
message, this must be `assistant_end`.
541
541
542
542
543
- This message indicates the conclusion of the assistant’ s response,
543
+ This message indicates the conclusion of the assistant' s response,
544
544
signaling that the assistant has finished speaking for the current
545
545
conversational turn.
546
546
custom_session_id :
@@ -560,8 +560,8 @@ types:
560
560
this must be `assistant_message`.
561
561
562
562
563
- This message contains both a transcript of the assistant’ s response
564
- and the expression measurement predictions of the assistant’ s audio
563
+ This message contains both a transcript of the assistant' s response
564
+ and the expression measurement predictions of the assistant' s audio
565
565
output.
566
566
custom_session_id :
567
567
type : optional<string>
@@ -860,7 +860,7 @@ types:
860
860
Parameters of the tool.
861
861
862
862
863
- These parameters define the inputs needed for the tool’ s execution,
863
+ These parameters define the inputs needed for the tool' s execution,
864
864
including the expected data type and description for each input field.
865
865
Structured as a stringified JSON schema, this format ensures the tool
866
866
receives data in the expected format.
@@ -912,7 +912,7 @@ types:
912
912
message, this must be `user_interruption`.
913
913
914
914
915
- This message indicates the user has interrupted the assistant’ s
915
+ This message indicates the user has interrupted the assistant' s
916
916
response. EVI detects the interruption in real-time and sends this
917
917
message to signal the interruption event. This message allows the
918
918
system to stop the current audio playback, clear the audio queue, and
@@ -937,7 +937,7 @@ types:
937
937
must be `user_message`.
938
938
939
939
940
- This message contains both a transcript of the user’ s input and the
940
+ This message contains both a transcript of the user' s input and the
941
941
expression measurement predictions if the input was sent as an [Audio
942
942
Input
943
943
message](/reference/empathic-voice-interface-evi/chat/chat#send.AudioInput.type).
@@ -1363,7 +1363,7 @@ types:
1363
1363
Tool.
1364
1364
1365
1365
1366
- These parameters define the inputs needed for the Tool’ s execution,
1366
+ These parameters define the inputs needed for the Tool' s execution,
1367
1367
including the expected data type and description for each input field.
1368
1368
Structured as a stringified JSON schema, this format ensures the tool
1369
1369
receives data in the expected format.
@@ -1410,17 +1410,15 @@ types:
1410
1410
text :
1411
1411
type : string
1412
1412
docs : >-
1413
- Instructions used to shape EVI’ s behavior, responses, and style.
1413
+ Instructions used to shape EVI' s behavior, responses, and style.
1414
1414
1415
1415
1416
1416
You can use the Prompt to define a specific goal or role for EVI,
1417
1417
specifying how it should act or what it should focus on during the
1418
1418
conversation. For example, EVI can be instructed to act as a customer
1419
1419
support representative, a fitness coach, or a travel advisor, each
1420
- with its own set of behaviors and response styles.
1421
-
1422
-
1423
- For help writing a system prompt, see our [Prompting
1420
+ with its own set of behaviors and response styles. For help writing a
1421
+ system prompt, see our [Prompting
1424
1422
Guide](/docs/speech-to-speech-evi/guides/prompting).
1425
1423
version :
1426
1424
type : integer
@@ -2142,7 +2140,7 @@ types:
2142
2140
The model temperature, with values between 0 to 1 (inclusive).
2143
2141
2144
2142
2145
- Controls the randomness of the LLM’ s output, with values closer to 0
2143
+ Controls the randomness of the LLM' s output, with values closer to 0
2146
2144
yielding focused, deterministic responses and values closer to 1
2147
2145
producing more creative, diverse responses.
2148
2146
source :
@@ -2293,6 +2291,11 @@ types:
2293
2291
Accepts a minimum value of 30 seconds and a maximum value of 1,800
2294
2292
seconds.
2295
2293
properties :
2294
+ duration_secs :
2295
+ type : optional<integer>
2296
+ docs : >-
2297
+ Duration in seconds for the timeout (e.g. 600 seconds represents 10
2298
+ minutes).
2296
2299
enabled :
2297
2300
type : boolean
2298
2301
docs : >-
@@ -2303,11 +2306,6 @@ types:
2303
2306
user inactivity being reached. However, the conversation will
2304
2307
eventually disconnect after 1,800 seconds (30 minutes), which is the
2305
2308
maximum WebSocket duration limit for EVI.
2306
- duration_secs :
2307
- type : optional<integer>
2308
- docs : >-
2309
- Duration in seconds for the timeout (e.g. 600 seconds represents 10
2310
- minutes).
2311
2309
source :
2312
2310
openapi : evi-openapi.json
2313
2311
inline : true
@@ -2321,6 +2319,11 @@ types:
2321
2319
Accepts a minimum value of 30 seconds and a maximum value of 1,800
2322
2320
seconds.
2323
2321
properties :
2322
+ duration_secs :
2323
+ type : optional<integer>
2324
+ docs : >-
2325
+ Duration in seconds for the timeout (e.g. 600 seconds represents 10
2326
+ minutes).
2324
2327
enabled :
2325
2328
type : boolean
2326
2329
docs : >-
@@ -2331,11 +2334,6 @@ types:
2331
2334
duration being reached. However, the conversation will eventually
2332
2335
disconnect after 1,800 seconds (30 minutes), which is the maximum
2333
2336
WebSocket duration limit for EVI.
2334
- duration_secs :
2335
- type : optional<integer>
2336
- docs : >-
2337
- Duration in seconds for the timeout (e.g. 600 seconds represents 10
2338
- minutes).
2339
2337
source :
2340
2338
openapi : evi-openapi.json
2341
2339
inline : true
@@ -2414,7 +2412,7 @@ types:
2414
2412
The model temperature, with values between 0 to 1 (inclusive).
2415
2413
2416
2414
2417
- Controls the randomness of the LLM’ s output, with values closer to 0
2415
+ Controls the randomness of the LLM' s output, with values closer to 0
2418
2416
yielding focused, deterministic responses and values closer to 1
2419
2417
producing more creative, diverse responses.
2420
2418
source :
0 commit comments