Code, refactor, and debug. One API for the coding tools you already use.
Let’s read the implementation and make a focused change.
src/utils/retry.tsexport async function withRetry<T>( task: () => Promise<T>, attempts = 3,): Promise<T> { for (let attempt = 0; ; attempt++) { try { return await task(); } catch (error) { if (attempt >= attempts - 1) throw error; const delay = 200 * 2 ** attempt; await new Promise(r => setTimeout(r, delay)); } }}Connect Codex, Cursor, Cline, and OpenCode. Stay in the environment you know.
https://xmodels.ai/api/coder/v1Write code, refactor modules, and build tests. Make every change easy to review.
@@ -8,3 +8,4 @@ 8 function label(name: string) { 9 − return name; 9 + const text = name.trim(); 10 + return text || 'Untitled'; 11 }
Bring context, files, and execution results together with function tool calling.
read_filesrc/utils/label.tsapply_patch+2 −1run_tests3 tests passedHandled empty input and covered three edge cases.
An OpenAI-compatible API, streaming responses, and straightforward token pricing.
model Xmodel-LLM-coder1.0 stream true tools [read_file, apply_patch]
Make your first Chat Completions request with your API key.
curl https://xmodels.ai/api/coder/v1/chat/completions \
-H "Authorization: Bearer $XMODELS_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"Xmodel-LLM-coder1.0",
"messages":[{"role":"user",
"content":"Hello, Xcoder."}]}'Use the OpenAI SDK. Just set the base URL and model.
from openai import OpenAI
import os
client = OpenAI(
base_url="https://xmodels.ai/api/coder/v1",
api_key=os.environ["XMODELS_API_KEY"],
max_retries=0,
)
response = client.chat.completions.create(
model="Xmodel-LLM-coder1.0",
messages=[{"role": "user", "content": "Hello, Xcoder."}],
)