API docs
OpenAI-compatible DeepSeek V4 Flash.
Use your Bypath key with /v1. Same OpenAI shape, built for coding agents.
Base URL
https://api.bypath.dev/v1
Authentication
Authorization: Bearer bp_live_...
Compatible routes
GET /v1/modelsPOST /v1/chat/completions
Model
deepseek-v4-flash
List models
curl https://api.bypath.dev/v1/models \
-H "Authorization: Bearer bp_live_YOUR_KEY"
Chat completion
curl https://api.bypath.dev/v1/chat/completions \
-H "Authorization: Bearer bp_live_YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-v4-flash",
"messages": [
{ "role": "user", "content": "Hello from Bypath" }
]
}'
Streaming
curl https://api.bypath.dev/v1/chat/completions \
-H "Authorization: Bearer bp_live_YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-v4-flash",
"stream": true,
"messages": [
{ "role": "user", "content": "Write a tiny TypeScript function" }
]
}'
OpenAI SDK
import OpenAI from "openai";
const client = new OpenAI({
apiKey: process.env.BYPATH_API_KEY,
baseURL: "https://api.bypath.dev/v1",
});
const response = await client.chat.completions.create({
model: "deepseek-v4-flash",
messages: [{ role: "user", content: "Hello" }],
});
console.log(response.choices[0]?.message?.content);
Guardrails
- 20 requests per minute by default.
- 1 in-flight request per API key by default.
- Prompt and response bodies are not stored by Bypath.
store: trueis rejected.