BUILT PRODUCTAI & Developer ToolsStudio-built
Building
ModelHub
One Key. Every Model. Pay in Rupees. Access hundreds of AI models with UPI top-ups, fallbacks, and spend limits.
Supported Models
100+ Leading LLMs
Payment Model
Instant UPI & Rupees
Compatibility
100% OpenAI Format
Product Capabilities & Architecture
10 Production FeaturesExhaustive System Features & Workflow
Unified API Core & UPI Rupee Billing
Drop-in OpenAI base_url endpoint requiring zero code changes in LangChain, LlamaIndex, or raw fetch calls
Seamless UPI top-ups via QR scan, dynamic VPA links, and optional auto-recharge thresholds
GST tax invoicing and monthly corporate expense statements downloadable in 1 click
Per-key spend tracking enabling granular budget isolation between staging, development, and production
Zero-Downtime Fallback & Smart Routing
Automated cascading failover across providers preventing outages when OpenAI or Anthropic suffers downtime
Latency-optimized routing directing requests to the fastest available regional edge endpoint
Cost-saver router automatically forwarding simple queries to ultra-cheap models (e.g. GPT-4o-mini / Haiku)
Custom provider priority ordering allowing developers to define explicit failover chains
Semantic Caching & Spend Management
Vector-similarity semantic response cache serving repeated queries with sub-20ms latency
Configurable similarity thresholds ensuring cache hits only return strictly equivalent responses
Team role-based access control with member invite links and individual developer budget allowances
Automated Slack and WhatsApp alerts when API keys consume 80% and 100% of allocated spend caps
Multi-Model Playground & Indic Analytics
Interactive web playground comparing outputs, tokens per second, and costs side-by-side in real time
Prompt history and evaluation testing suite measuring regression across model version upgrades
Indic language benchmark leaderboard measuring reasoning quality across Hindi, Hinglish, and regional scripts
Exportable Python and cURL code snippets configured with active API keys ready for production
All Individual Feature Specifications:
01One Universal API Key: Standard OpenAI-compatible API format (drop-in base_url replacement) routing to GPT-4o, Claude 3.5, Gemini 1.5, and Llama 3.1
02Pay in Rupees via UPI: Top up balances instantly via PhonePe, Google Pay, or Paytm without requiring international credit cards or paying forex markups
03Zero-Downtime Smart Fallbacks: Automatically reroutes traffic to alternative models or providers (e.g. Claude to Gemini) if a primary provider experiences downtime or rate limits
04Semantic Response Caching: Caches frequent queries with vector similarity, reducing API latency to under 20ms and slashing token bills by up to 40%
05Team Workspaces & Hard Spend Limits: Create scoped API keys with strict budget ceilings ($5 or ₹500 max) to prevent rogue loops or unexpected runaway bills
06Side-by-Side Model Playground: Compare speed, quality, and token costs across 4 models simultaneously in real time
07Prompt Optimization & Token Trimmer: Automatically strips redundant whitespace, system noise, and trailing tokens to minimize prompt costs
08Detailed Cost & Latency Analytics: Real-time visual dashboard breaking down token usage, cost per request, error rates, and latency percentiles
09India-First Regional Language Routing: Routes Indic-language prompts to models fine-tuned for Hindi, Tamil, Telugu, and Marathi with optimized tokenizers
10Self-Host Gateway Binary: Open-source local proxy developers can run inside private VPCs with strict zero-logging data privacy guarantees
Visual Interface Showcase
Screenshots & Interaction Design
High-resolution interface viewports ready to link direct screenshots and walkthrough captures.
https://modelhub.app
DESKTOP VIEWActive Production View: ModelHub
FEATURE DETAILVIEWPORT 02
Interactive Workspace ViewSlot for modal, editor, or action flow preview.
MOBILE ADAPTIVEVIEWPORT 03
Mobile & Tablet ResponsiveSlot for touch-screen mobile viewport screenshot.
Who experiences this problem?
Indie hackers, college students, software engineering teams, SaaS founders, enterprise architects, and digital marketing agencies across India building AI features without corporate credit cards.
Current Workflow Breakdown:Card declines on OpenAI/Anthropic due to RBI e-mandate rules, 3.5% foreign exchange fees, sudden provider outages breaking production apps, and managing disparate API SDKs.
How our product addresses it
A unified AI model gateway and intelligent router: developers get a single OpenAI-compatible API key to access 100+ state-of-the-art LLMs (OpenAI, Anthropic Claude, Google Gemini, Meta Llama, Mistral) with instant UPI and rupee payments, automatic zero-downtime fallbacks, semantic caching, and team spend limits.
Technical Stack & Foundations
Cloudflare Workers Edge Proxy Core
OpenAI-Compatible /v1 API Endpoint
Smart Fallback & Latency-Aware Model Router
Semantic Vector Caching with Redis & Qdrant
Razorpay UPI Top-Up & Rupee Billing Ledger
Streaming SSE Token Multiplexer
Granular Token Rate Limiting & Spend Caps
Interactive Multi-Model Side-by-Side Playground
Current progress and roadmap
Stage: Building
Deploying the ultra-low latency edge gateway proxy (Cloudflare Workers), testing automatic model failover routing, and integrating Razorpay dynamic UPI autopay top-ups.
Other studio initiatives
LeadPilot
An AI sales assistant that finds leads, writes your pitches, replies on WhatsApp and writes notes after every call.
Multi-Source Lead Radar: Continuously scans Upwork, Fiverr, LinkedIn Jobs, RemoteOK, Reddit, X (Twitter), Google Maps, and public business directories for live hiring intent
AI 0-100 Lead Scoring: Rates every lead based on skill alignment, budget readiness, urgency, and reply probability so you only prioritize high-yield opportunities
Zero-Ban 'Draft & Approve' Guard: Generates tailored proposals and pitches for user 1-click review, preventing automated bot detection and protecting platform accounts from bans
+6 more technical capabilities
Studio-built
Kaam.ai
An AI agent that fills forms, handles files and runs tasks in your browser and terminal, and asks before anything risky.
Autonomous Form Filling: Navigates multi-page application forms, auto-mapping user profile data into input fields with pixel-level precision
Risk-Aware Approval Gates: Pauses before critical actions like 'Submit Form', 'Delete Files', or 'Execute Payment', requesting clear user confirmation
Kaam Terminal CLI: Trigger and run entire automated workflows directly from your terminal using commands like `kaam run admission-form`
+4 more technical capabilities
Studio-built