Skip to content

Commit 9c89456

Browse files
goosewingoosewin
committed
feat: break down chat doc into 4 separate docs
1 parent 3155e88 commit 9c89456

File tree

4 files changed

+1422
-0
lines changed

4 files changed

+1422
-0
lines changed

‎fern/chat/non-streaming.mdx‎

Lines changed: 356 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,356 @@
1+
---
2+
title: Non-streaming chat
3+
subtitle: Simple request-response chat patterns with complete responses
4+
---
5+
6+
## Overview
7+
8+
Non-streaming chat returns the complete response after the assistant finishes processing. This is ideal for simple integrations where you don't need real-time output.
9+
10+
Benefits:
11+
- **Simple integration** - Single request, single response
12+
- **Complete responses** - Get the full message at once
13+
- **Lower complexity** - No need to handle streaming events
14+
- **Better for batch processing** - Process multiple requests sequentially
15+
16+
## Basic example
17+
18+
```bash
19+
curl -X POST https://api.vapi.ai/chat \
20+
-H "Authorization: Bearer YOUR_API_KEY" \
21+
-H "Content-Type: application/json" \
22+
-d '{
23+
"assistantId": "your-assistant-id",
24+
"input": "What is the weather like today?"
25+
}'
26+
```
27+
28+
### JavaScript implementation
29+
30+
```javascript
31+
const response = await fetch('https://api.vapi.ai/chat', {
32+
method: 'POST',
33+
headers: {
34+
'Authorization': 'Bearer YOUR_API_KEY',
35+
'Content-Type': 'application/json'
36+
},
37+
body: JSON.stringify({
38+
assistantId: 'your-assistant-id',
39+
input: 'What is the weather like today?'
40+
})
41+
});
42+
43+
const chat = await response.json();
44+
console.log(chat.output[0].content); // Assistant's response
45+
```
46+
47+
## Response structure
48+
49+
```javascript
50+
{
51+
"id": "chat_123456",
52+
"orgId": "org_789012",
53+
"assistantId": "assistant_345678",
54+
"name": "Weather Chat",
55+
"sessionId": "session_901234",
56+
"messages": [
57+
{
58+
"role": "user",
59+
"content": "What is the weather like today?"
60+
}
61+
],
62+
"output": [
63+
{
64+
"role": "assistant",
65+
"content": "I'd be happy to help with weather information, but I'll need to know your location first. What city are you in?"
66+
}
67+
],
68+
"createdAt": "2024-01-15T09:30:00Z",
69+
"updatedAt": "2024-01-15T09:30:01Z"
70+
}
71+
```
72+
73+
## Custom assistant configuration
74+
75+
Pass assistant configuration inline instead of using a pre-created assistant:
76+
77+
```bash
78+
curl -X POST https://api.vapi.ai/chat \
79+
-H "Authorization: Bearer YOUR_API_KEY" \
80+
-H "Content-Type: application/json" \
81+
-d '{
82+
"input": "Help me plan a trip to Paris",
83+
"assistant": {
84+
"model": {
85+
"provider": "openai",
86+
"model": "gpt-4",
87+
"temperature": 0.7,
88+
"messages": [
89+
{
90+
"role": "system",
91+
"content": "You are a helpful travel assistant specializing in European destinations."
92+
}
93+
]
94+
},
95+
"voice": {
96+
"provider": "azure",
97+
"voiceId": "andrew"
98+
}
99+
}
100+
}'
101+
```
102+
103+
### JavaScript with custom assistant
104+
105+
```javascript
106+
const response = await fetch('https://api.vapi.ai/chat', {
107+
method: 'POST',
108+
headers: {
109+
'Authorization': 'Bearer YOUR_API_KEY',
110+
'Content-Type': 'application/json'
111+
},
112+
body: JSON.stringify({
113+
input: 'Help me plan a trip to Paris',
114+
assistant: {
115+
model: {
116+
provider: 'openai',
117+
model: 'gpt-4',
118+
temperature: 0.7,
119+
messages: [
120+
{
121+
role: 'system',
122+
content: 'You are a helpful travel assistant specializing in European destinations.'
123+
}
124+
]
125+
},
126+
voice: {
127+
provider: 'azure',
128+
voiceId: 'andrew'
129+
}
130+
}
131+
})
132+
});
133+
134+
const chat = await response.json();
135+
console.log(chat.output[0].content);
136+
```
137+
138+
## Context management
139+
140+
### Using sessions
141+
142+
Sessions allow multiple chats to share the same context:
143+
144+
```bash
145+
# Create a session first
146+
curl -X POST https://api.vapi.ai/session \
147+
-H "Authorization: Bearer YOUR_API_KEY" \
148+
-H "Content-Type: application/json" \
149+
-d '{
150+
"assistantId": "your-assistant-id"
151+
}'
152+
```
153+
154+
```bash
155+
# Use the session for multiple chats
156+
curl -X POST https://api.vapi.ai/chat \
157+
-H "Authorization: Bearer YOUR_API_KEY" \
158+
-H "Content-Type: application/json" \
159+
-d '{
160+
"sessionId": "session_abc123",
161+
"input": "My name is Alice"
162+
}'
163+
```
164+
165+
```bash
166+
# Continue the conversation
167+
curl -X POST https://api.vapi.ai/chat \
168+
-H "Authorization: Bearer YOUR_API_KEY" \
169+
-H "Content-Type: application/json" \
170+
-d '{
171+
"sessionId": "session_abc123",
172+
"input": "What is my name?"
173+
}'
174+
```
175+
176+
### JavaScript session example
177+
178+
```javascript
179+
// Create a session
180+
const sessionResponse = await fetch('https://api.vapi.ai/session', {
181+
method: 'POST',
182+
headers: {
183+
'Authorization': 'Bearer YOUR_API_KEY',
184+
'Content-Type': 'application/json'
185+
},
186+
body: JSON.stringify({
187+
assistantId: 'your-assistant-id'
188+
})
189+
});
190+
191+
const session = await sessionResponse.json();
192+
193+
// Use the session for multiple chats
194+
const chat1 = await fetch('https://api.vapi.ai/chat', {
195+
method: 'POST',
196+
headers: {
197+
'Authorization': 'Bearer YOUR_API_KEY',
198+
'Content-Type': 'application/json'
199+
},
200+
body: JSON.stringify({
201+
sessionId: session.id,
202+
input: 'My name is Alice'
203+
})
204+
});
205+
206+
const chat1Data = await chat1.json();
207+
console.log(chat1Data.output[0].content); // "Nice to meet you, Alice!"
208+
209+
const chat2 = await fetch('https://api.vapi.ai/chat', {
210+
method: 'POST',
211+
headers: {
212+
'Authorization': 'Bearer YOUR_API_KEY',
213+
'Content-Type': 'application/json'
214+
},
215+
body: JSON.stringify({
216+
sessionId: session.id,
217+
input: 'What is my name?' // Assistant will remember "Alice"
218+
})
219+
});
220+
221+
const chat2Data = await chat2.json();
222+
console.log(chat2Data.output[0].content); // "Your name is Alice."
223+
```
224+
225+
### Using previous chat
226+
227+
Link chats together without creating a session:
228+
229+
```bash
230+
# First chat
231+
curl -X POST https://api.vapi.ai/chat \
232+
-H "Authorization: Bearer YOUR_API_KEY" \
233+
-H "Content-Type: application/json" \
234+
-d '{
235+
"assistantId": "your-assistant-id",
236+
"input": "I need help with Python programming"
237+
}'
238+
```
239+
240+
```bash
241+
# Continue conversation with previous chat ID
242+
curl -X POST https://api.vapi.ai/chat \
243+
-H "Authorization: Bearer YOUR_API_KEY" \
244+
-H "Content-Type: application/json" \
245+
-d '{
246+
"assistantId": "your-assistant-id",
247+
"previousChatId": "chat_abc123",
248+
"input": "Show me how to create a list"
249+
}'
250+
```
251+
252+
### JavaScript previous chat example
253+
254+
```javascript
255+
// First chat
256+
const firstChat = await fetch('https://api.vapi.ai/chat', {
257+
method: 'POST',
258+
headers: {
259+
'Authorization': 'Bearer YOUR_API_KEY',
260+
'Content-Type': 'application/json'
261+
},
262+
body: JSON.stringify({
263+
assistantId: 'your-assistant-id',
264+
input: 'I need help with Python programming'
265+
})
266+
});
267+
268+
const firstChatData = await firstChat.json();
269+
270+
// Continue conversation
271+
const secondChat = await fetch('https://api.vapi.ai/chat', {
272+
method: 'POST',
273+
headers: {
274+
'Authorization': 'Bearer YOUR_API_KEY',
275+
'Content-Type': 'application/json'
276+
},
277+
body: JSON.stringify({
278+
assistantId: 'your-assistant-id',
279+
previousChatId: firstChatData.id,
280+
input: 'Show me how to create a list'
281+
})
282+
});
283+
284+
const secondChatData = await secondChat.json();
285+
console.log(secondChatData.output[0].content);
286+
```
287+
288+
## Message types
289+
290+
### User message
291+
```javascript
292+
{
293+
"role": "user",
294+
"content": "Hello, how are you?",
295+
"name": "Alice" // Optional
296+
}
297+
```
298+
299+
### Assistant message
300+
```javascript
301+
{
302+
"role": "assistant",
303+
"content": "I'm doing well, thank you!",
304+
"tool_calls": [...], // Optional tool calls
305+
"refusal": null // Optional refusal message
306+
}
307+
```
308+
309+
### System message
310+
```javascript
311+
{
312+
"role": "system",
313+
"content": "You are a helpful assistant specializing in technical support."
314+
}
315+
```
316+
317+
### Tool message
318+
```javascript
319+
{
320+
"role": "tool",
321+
"content": "Weather data retrieved successfully",
322+
"tool_call_id": "call_123456"
323+
}
324+
```
325+
326+
### Developer message
327+
```javascript
328+
{
329+
"role": "developer",
330+
"content": "Always be concise in your responses."
331+
}
332+
```
333+
334+
## Best practices
335+
336+
<Tip>
337+
**When to use non-streaming**
338+
- Simple chatbots or FAQ systems
339+
- Batch processing of messages
340+
- When you need the complete response before proceeding
341+
- Integrations where real-time display isn't important
342+
</Tip>
343+
344+
<Note>
345+
**Performance considerations**
346+
- Responses may take longer for complex queries
347+
- Use streaming for better user experience with long responses
348+
- Consider timeout settings for your requests
349+
</Note>
350+
351+
## Next steps
352+
353+
- Try [streaming chat](/docs/chat/streaming) for real-time responses
354+
- Explore [OpenAI compatibility](/docs/chat/openai-compatibility) for existing integrations
355+
- View the [API reference](/api-reference/chats/chat-controller-create-chat) for all parameters
356+
- Learn about [assistants](/docs/assistants) to customize behavior

0 commit comments

Comments
 (0)