One SDK. Smart routing across OpenAI, Anthropic, Gemini, and Llama via Groq. Automatic failover. Cost optimisation. Zero lock-in.
from uniq import AI ai = AI(api_key="uq_••••••••••••••••") response = ai.complete( "Explain transformer architecture", model="auto", # routes to best provider) print(response.text) # ✓ Provider : anthropic → claude-sonnet-4-6# ✓ Latency : 1.24s │ Tokens: 342 │ $0.0014Platform
Built for developers who want the simplest possible interface to the most powerful AI models — without sacrificing control.
Automatically classifies your prompt — code, math, analysis, translation — and routes to the model with the strongest performance for that task. No config required.
If a provider is down, rate-limiting, or times out, the request is instantly retried against the next healthy provider in your fallback chain — fully transparent to your app.
Routes to the cheapest model capable of handling the request. The routing engine balances latency, quality, and per-token cost in real time across all configured providers.
Every call logged with provider, model, token counts, latency, and cost. Per-request traces, hourly latency percentiles, and live charts in the dashboard.
Generate scoped uq_* keys with RPM limits, daily token quotas, and monthly spend caps. Revoke compromised credentials instantly without redeploying.
Typed Python SDK with sync and async clients, streaming, tool calling, and exponential-backoff retries baked in. Drop-in compatible with the OpenAI client format.
Ingest documents, search with BM25 ranking, and ground LLM answers in your own data. The /rag CLI command indexes files in seconds — no vector DB required.
Run MMLU, HumanEval, and GSM8K suites against any provider/model combination. CI regression checks and a public leaderboard included out of the box.
Schedule recurring LLM tasks — summaries, reports, alerts — using a simple cron expression. Results are stored and accessible via API or the dashboard.
Create organisations, invite members, assign roles, and enforce per-org monthly budget limits. SSO via OIDC/SAML for enterprise teams.
Inline AI completion straight in your editor via the VS Code extension. MCP server exposes the full API to Claude Desktop and any MCP-compatible client.
Add the uniq bot to any Slack channel. Multi-turn conversations, file uploads, and slash commands — all routed through your configured providers and keys.
Pricing
Start free. Scale when you're ready. No hidden fees, no egress charges.
For side projects and exploration.
For developers shipping real products.
For teams with production workloads.
For large-scale, compliance-critical deployments.
All plans include API key management and the full Python SDK. Prices in USD. Cancel anytime.