8 models across 4 providers. Every response passes through VASVELVOGVEG — quality-scored, safety-gated, and recorded to an immutable ledger. One API call. Any model. Zero lock-in.
Fast general-purpose model. Excellent for chat, summarization, and lightweight reasoning tasks.
Advanced reasoning model with chain-of-thought. Excels at math, logic, and multi-step analysis.
Multimodal vision-language model. Analyzes images, screenshots, and documents with natural language.
High-performance text embeddings for semantic search, RAG pipelines, and document clustering.
Google's fastest multimodal model. Exceptional at code, analysis, and long-context tasks.
OpenAI's flagship multimodal model. Best-in-class for complex reasoning and creative tasks.
Smaller, faster, cheaper variant of GPT-4o. Great for high-volume applications.
ZeniBot's ethical fallback. When all providers are down, Dharma responds from a curated Tamil philosophical corpus. ZeniBot is never silent.
Unlike raw API calls, ZeniBot wraps every model in a 6-layer intelligence pipeline.
Six configurable compliance filters: No Harm, Truth, Consent, Transparency, Restorative-First, Service-Orientation.
PID controller adjusts quality scores with EWMA drift detection. Not just vibes — real signal processing.
Every request → evaluation → response is recorded in an immutable blockchain-style ledger. Full audit trail.
Circuit breaker detects provider failures. Automatic failover across 4 providers — your app never knows.
import requests response = requests.post( "https://zenibot.dev/api/generate", headers={"Authorization": "Bearer zb_your_key"}, json={"prompt": "Explain quantum computing"} ) data = response.json() print(data["response"]) # AI response print(data["provider"]) # Which model served it print(data["alignment"]) # Quality + safety scores print(data["ledger_block"]) # Immutable audit trail
Free forever · 100 requests/day · No credit card