Question 1

Do you support Indian payment methods and GST invoices?

Accepted Answer

Yes! We accept Indian Credit/Debit Cards, UPI, Netbanking, and corporate wire transfers. All Indian business customers receive automated GST-compliant tax invoices with their GSTIN for seamless input tax credit.

Question 2

How does Bring Your Own Key (BYOK) work?

Accepted Answer

When you use BYOK, you simply paste your existing API keys for Anthropic, OpenAI, Google Cloud, or Mistral into your encrypted dashboard. Your requests are proxied directly to providers with zero token markup. You only pay v43.ai our low monthly platform subscription for intelligent routing, fallbacks, and evaluation.

Question 3

What happens if I exceed my monthly router token quota?

Accepted Answer

If you're using BYOK, your models will continue executing uninterrupted. For platform-managed tokens, overages are billed at a simple flat rate of ₹200 per 1M router tokens on Pro, and ₹150 per 1M on Scale / Team. You can also configure hard budget caps in your dashboard to prevent unexpected charges.

Question 4

Can I switch between monthly and annual billing?

Accepted Answer

Yes. You can toggle between monthly and annual billing at any time directly from the billing portal. Annual plans include an upfront 20% discount. Upgrades take effect immediately with prorated billing.

Question 5

How does v43.ai achieve 40%+ cost savings on LLM requests?

Accepted Answer

Most production apps send all prompts to top-tier reasoning models like Claude 3.7 or GPT-4o, even for simple classification, extraction, or formatting tasks. v43.ai's dynamic router analyzes prompt complexity and context size in real time. It automatically routes straightforward queries to ultra-fast, cost-effective models (like Gemini 2.0 Flash or local Llama) while routing complex code generation to frontier reasoning models.

Question 6

Can I connect private or self-hosted models running on Ollama or vLLM in India?

Accepted Answer

Yes! v43.ai provides a lightweight secure agent daemon that exposes your local Ollama or private GPU vLLM cluster to your orchestration routing pool. This enables hybrid deployments where sensitive data stays on-premise in India while non-sensitive tasks utilize cloud models.

Question 7

Do you offer private VPC deployments in Indian Cloud Regions (AWS Mumbai, GCP Delhi)?

Accepted Answer

Yes. Enterprise customers can deploy the v43.ai orchestration plane directly inside their AWS Mumbai (ap-south-1), Google Cloud Mumbai/Delhi, or Azure India VPC using our production Terraform and Helm modules, ensuring strict local data residency.

Question 8

Is customer data stored or used to train models?

Accepted Answer

Never. v43.ai acts as an in-memory streaming proxy. We do not store prompt inputs or outputs unless you explicitly configure payload tracing in your dashboard. Enterprise plans come with a formal Zero-Data Retention (ZDR) guarantee and DPDP Act compliance.

Features & Capabilities	Basic	Pro	Scale / Team	Enterprise
Orchestration & Routing
Dynamic Multi-Model Router Intelligently routes prompts to the best model based on latency, cost, and task capability.	Standard Router	Autonomous AI Router	Autonomous AI Router	Custom Trained Policy Router
Automatic Fallback Cascades Zero-error failovers if an upstream provider experiences downtime or rate limits.
Parallel Consensus Voting Broadcast prompts to multiple models concurrently and synthesize the highest-scored answer.		Up to 3 models	Up to 8 models	Unlimited models
Bring Your Own Keys (BYOK) Plug in your own OpenAI, Anthropic, Google, and Mistral API keys without commission.
Local LLM Bridge (Ollama / vLLM) Route requests seamlessly between private local hardware and cloud models.
Routing Overhead Latency Gateway proxy time added to raw upstream model latency.	<80ms	<40ms	<20ms	<5ms (Dedicated Edge)
Model Access & Limits
Included Router Tokens Managed token quota included in base monthly subscription.	100K / mo	1.5M / mo	6M / mo	Custom Pool
Concurrent In-flight Requests Simultaneous open model connections.	5	25	100	Custom Dedicated
Supported Foundation Models Access to latest Claude, Gemini, GPT-4o, DeepSeek, Llama models out of the box.	15+ Models	50+ Models	50+ Models	All Models + Custom Models
Custom Fine-tuned Endpoints Connect proprietary internal fine-tunes or HuggingFace endpoints.
Analytics, Tracing & Quality
Execution History & Tracing Inspection window for request payloads, prompts, tokens, and latency traces.	14 days	30 days	90 days	Unlimited / S3 Archival
Cost Allocation per User/Tenant Attribute token spend and cost metrics directly to specific end-users or API keys.
Prompt Regression Testing Suite Run golden dataset benchmarks before updating system prompts or model versions.		Manual Runner	Automated CI/CD	Full Golden Suite + Webhooks
OpenTelemetry Export Stream traces directly to Datadog, New Relic, or Prometheus.
Security, Governance & Support
Data Privacy & Zero Data Retention Strict policy guaranteeing customer prompt data is never stored or used to train models.	Standard Privacy	Standard Privacy	ZDR Available	Guaranteed Zero Retention + BAA
Team Seats Collaborators and dashboard accounts.	1 user	1 user	10 seats included	Unlimited
Compliance & Invoicing Verified security certifications, GST compliance and auditor access.	GST Invoicing	GST Invoicing	GST + DPDP Compliant	ISO 27001, SOC 2, HIPAA, DPDP
Deployment Topology Where the orchestration routing proxy runs.	Multi-tenant Cloud	Multi-tenant Cloud	Priority Multi-tenant	Dedicated VPC (India / Global)
Support SLA Response time and support availability.	Email & Discord	Priority Email (4h)	Shared Slack (2h)	Dedicated TAM & 15m Critical SLA

Simple, Predictable Plans for Every AI Workload

Basic

Pro

Scale / Team

Enterprise

Calculate Your Token Savings

Compare All Plan Features

Got Questions? We Have Answers

Ready to Orchestrate at Peak Performance?