Governed AI calls in minutes.

Tokenality is invitation-only during stealth. Request design-partner access and we'll provision your organization — an org is created and a per-org encryption key is generated on first sign-in. You'll receive your gateway base URL at onboarding (it looks like https://gateway.tokenality.ai).

Sign in at app.tokenality.ai with SSO. From the Tokens page, issue a virtual token key scoped to a project and budget. It looks like tk_live_… — this is what your code uses, not your provider key.

Point your existing SDK at the gateway and use your virtual key. Your code doesn't change — just the base URL and the key.

# OpenAI SDK — Python
from openai import OpenAI

client = OpenAI(
    base_url="https://gateway.tokenality.ai/v1/openai/v1",
    api_key="tk_live_…",   # your Tokenality virtual key
)
resp = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello"}],
)

# Anthropic — curl
curl https://gateway.tokenality.ai/v1/anthropic/v1/messages \
  -H "authorization: Bearer tk_live_…" \
  -H "content-type: application/json" \
  -d '{"model":"claude-sonnet-4-6","max_tokens":100,
       "messages":[{"role":"user","content":"Hello"}]}'

Every call passes the Token Firewall (budget + policy + PII pre-flight) before it reaches the provider, then lands as one row in the Token Ledger — reconciled to the person, project, and task, with exact cost. If a call would breach the budget, the gateway returns 402 before spending a token.

Add hard budgets per key, bring your own provider keys (BYOK — your keys are encrypted at rest and bound to your org), turn on PII fail-closed, and invite your team. Governance is on by default at every tier.