Omni Docs
Start Here

Quickstart

Generate your API key, make your first authenticated call, and run an agent session.

Get started with Omni in three steps using your developer API key.

Prerequisites

RequirementSourceDescription
API KeyDeveloper ConsoleKey prefixed with omni_sk_live_
Base URLhttps://edge.omnistatic.comPrimary developer API origin

Export your API key into your terminal environment:

export OMNI_API_KEY="omni_sk_live_..."
export OMNI_BASE_URL="https://edge.omnistatic.com"

1. Verify Your Credential

Verify your token and inspect your account identity:

curl -sS "$OMNI_BASE_URL/v1/me" \
  -H "Authorization: Bearer $OMNI_API_KEY"
const res = await fetch("https://edge.omnistatic.com/v1/me", {
  headers: {
    Authorization: `Bearer ${process.env.OMNI_API_KEY}`,
  },
});
const data = await res.json();
console.log("Authenticated as:", data.userId);
import os
import requests

res = requests.get(
    "https://edge.omnistatic.com/v1/me",
    headers={"Authorization": f"Bearer {os.environ['OMNI_API_KEY']}"}
)
print("Authenticated as:", res.json().get("userId"))

Response:

{
  "ok": true,
  "userId": "usr_94a645d99c3a",
  "profile": {
    "userName": "developer",
    "display_name": "Dev User"
  },
  "verifiedAt": "2026-08-16T00:00:00.000Z"
}

2. Store a Semantic Memory

Save a piece of knowledge into your agent's persistent memory bank:

curl -sS -X POST "$OMNI_BASE_URL/v1/memories" \
  -H "Authorization: Bearer $OMNI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "content": "Project Phoenix launch is scheduled for September 1st, 2026."
  }'
const res = await fetch("https://edge.omnistatic.com/v1/memories", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.OMNI_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    content: "Project Phoenix launch is scheduled for September 1st, 2026.",
  }),
});
const memory = await res.json();
console.log("Memory saved with ID:", memory.id);
import os
import requests

res = requests.post(
    "https://edge.omnistatic.com/v1/memories",
    headers={
        "Authorization": f"Bearer {os.environ['OMNI_API_KEY']}",
        "Content-Type": "application/json"
    },
    json={"content": "Project Phoenix launch is scheduled for September 1st, 2026."}
)
print("Memory ID:", res.json().get("id"))

Response:

{
  "id": "mem_01jze9m2xw",
  "ownerId": "usr_94a645d99c3a",
  "content": "Project Phoenix launch is scheduled for September 1st, 2026.",
  "createdAt": "2026-08-16T00:00:00.000Z"
}

3. Run an Agent Chat Session

Send a prompt to the model. Relevant memories are automatically retrieved and injected into the context window:

curl -sS -X POST "$OMNI_BASE_URL/v1/chat/completions" \
  -H "Authorization: Bearer $OMNI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini",
    "messages": [
      { "role": "user", "content": "When is the launch of Project Phoenix?" }
    ]
  }'
const res = await fetch("https://edge.omnistatic.com/v1/chat/completions", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.OMNI_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "gpt-4o-mini",
    messages: [
      { role: "user", content: "When is the launch of Project Phoenix?" },
    ],
  }),
});

const reader = res.body?.getReader();
const decoder = new TextDecoder();
while (reader) {
  const { done, value } = await reader.read();
  if (done) break;
  console.log(decoder.decode(value));
}
import os
import requests

res = requests.post(
    "https://edge.omnistatic.com/v1/chat/completions",
    headers={
        "Authorization": f"Bearer {os.environ['OMNI_API_KEY']}",
        "Content-Type": "application/json"
    },
    json={
        "model": "gpt-4o-mini",
        "messages": [{"role": "user", "content": "When is the launch of Project Phoenix?"}]
    },
    stream=True
)

for line in res.iter_lines():
    if line:
        print(line.decode('utf-8'))

The agent inspects memory, finds the fact, and replies with index-grounded certainty.


Troubleshooting & Failure Modes

Status CodeCauseRecommended Action
401 UnauthorizedInvalid key or missing Bearer headerVerify that OMNI_API_KEY is exported and valid.
403 ForbiddenKey lacks required scope or account not in allowlistCheck scopes at /dev/keys or ensure key has chat and memory:read.
429 Too Many RequestsExceeded 60 req/min key limit or 600 req/hr UID limitImplement exponential backoff retry.
429 Quota ExceededMonthly token allowance or credit ceiling reachedInspect current consumption via GET /v1/credits.
500 Upstream ErrorUpstream model provider temporarily unreachableRetry with backoff using the same payload.

Coding Agent Instruction Block

Paste this prompt into Claude Code, Cursor, or Codex to have an assistant set up your integration:

Connect to the Omni Developer API at https://edge.omnistatic.com using the Bearer token in OMNI_API_KEY.
1. Verify connectivity with GET /v1/me.
2. Ingest contextual facts using POST /v1/memories {"content": "<text>"}.
3. Execute chat completions via POST /v1/chat/completions with standard OpenAI message payloads.
Handle 429 rate limits with exponential backoff.

On this page