Coding Agents
VectorAxis is an OpenAI-compatible gateway, so any AI coding tool that accepts a custom base_url routes through it with a single configuration change — no plugin, no wrapper. Your whole engineering team gets cost tracking, provider fallbacks, caching, budgets, and guardrails without changing how the tools work.
Because the gateway is OpenAI-compatible, you send OpenAI-format chat requests and VectorAxis translates them to whatever provider the key points at. So a coding tool that only speaks “OpenAI” can still use Anthropic, Groq, and more: create a virtual key for that provider, pick a model it serves (e.g. claude-haiku-4-5), and VectorAxis handles the format under the hood — no separate Anthropic endpoint needed.
| Tool | How to point it at VectorAxis |
|---|---|
| Cursor | Settings → Models → OpenAI API Key: enable “Override OpenAI Base URL”, set it to https://api.vectoraxis.ai/v1, and paste a virtual key as the API key. |
| Continue.dev | In ~/.continue/config.json, set each model’s apiBase and apiKey (see example below). |
| Cody (Sourcegraph) | Add a custom OpenAI-compatible provider in Cody settings with the VectorAxis base URL and a virtual key. |
| Aider | Set OPENAI_API_BASE=https://api.vectoraxis.ai/v1 and OPENAI_API_KEY=vk-… (or the matching CLI flags). |
| Any OpenAI SDK / tool | Set base_url + api_key (a virtual key slug). |
from openai import OpenAI
client = OpenAI(
base_url="https://api.vectoraxis.ai/v1",
api_key="vk-1a2b3c4d5e6f7a8b", # your VectorAxis virtual key
)
# model + everything else stays the same
client.chat.completions.create(
model="gpt-4o",
messages=[{"role": "user", "content": "Explain this function"}],
)Two models, each behind its own virtual key — a GPT key and a Claude key. Both use the openai provider type (that just means “OpenAI-compatible protocol”); VectorAxis routes to the real provider based on the key.
{
"models": [
{
"title": "VectorAxis -> GPT-4o",
"provider": "openai",
"model": "gpt-4o",
"apiBase": "https://api.vectoraxis.ai/v1",
"apiKey": "vk-1a2b3c4d5e6f7a8b"
},
{
"title": "VectorAxis -> Claude",
"provider": "openai",
"model": "claude-haiku-4-5",
"apiBase": "https://api.vectoraxis.ai/v1",
"apiKey": "vk-c7d4e1f2a3b95a8b"
}
]
}| Capability | How it helps with coding agents |
|---|---|
| Cost visibility | See exactly what each developer, team, or project spends on AI, broken down by model and day, in Analytics. |
| Provider fallback | If a provider has a bad minute mid-session, retry & fallback keeps the IDE responding instead of erroring out. |
| Caching | Repeated prompts (“explain this function”, autocomplete) can be served from the cache, cutting latency and cost. |
| Budget controls | Give each dev or team a virtual key with a credit limit and rate limits, so no one can blow the monthly budget. |
| Guardrails | Block or redact secrets and PII in a snippet before it ever leaves your network with guardrails. |
| Model flexibility | Send cheap tasks to a small model and hard ones to a frontier model with a routing config — no tool change. |