Frequently Asked Questions

UMVA AI is a pay-as-you-go AI platform that provides the cheapest AI inference for coding and general use. It includes an AI-powered IDE that runs on macOS, Windows, and Linux, along with API access for integrating AI into your own tools and workflows. All text models feature a 1M context window, and you only pay for what you use — no subscriptions, no monthly commitments.

UMVA AI stands out for its radically simple pay-as-you-go pricing with no subscriptions, a massive 1 million token context window on all text models, support for up to 384K output tokens, and a fully OpenAI-compatible API. Our IDE is completely free to download and use, and we accept both card and cryptocurrency payments without requiring KYC.

UMVA AI is pay-as-you-go from the start — no free trial needed because you only pay for what you use. Just add as little as $0.10 to your wallet and start coding immediately. No credit card is required to create an account.

Yes, you need to create a free UMVA account to use our AI services. Registration requires only an email address and password. No KYC or identity verification is required.

UMVA AI supports virtually every programming language and framework. Our models are trained on vast code corpora including Python, JavaScript, TypeScript, Go, Rust, C/C++, Java, Ruby, PHP, Swift, Kotlin, and hundreds more. The IDE provides syntax highlighting and language-aware completions for all major languages.

Absolutely. While UMVA AI is optimized for coding, our models excel at a wide range of tasks including document analysis, data extraction, summarization, translation, brainstorming, writing, research, and general reasoning. The 1M context window makes it particularly powerful for processing long documents and large datasets.

You add funds to your UMVA wallet and pay per token as you use AI models. Pricing is based on input and output tokens per million tokens — starting from as low as $0.50. Your usage is deducted from your balance in real time. There are no monthly fees, no subscription tiers, and no expiration on your credits.

Our pricing starts at $0.50 per million input tokens for umva-code-fast. umva-code-agent is also $0.50/M input tokens, umva-code-pro is $1.50/M input tokens, and umva-code-architect is $1.50/M input tokens. Output tokens are priced at $1.12/M for fast models and $3.30/M for premium models. All prices are in USD.

Topping up your wallet is simple. Log into your UMVA account, navigate to the Billing or Wallet section, choose an amount (minimum $1.00 for card or crypto), select your payment method, and complete the transaction. Funds are available instantly for card and crypto payments.

The minimum deposit is $1.00 for card and cryptocurrency payments. For bank transfers, the minimum is $10.00. There is no maximum deposit limit, so you can add as much as you like for convenience.

No. Funds in your UMVA wallet never expire. You can keep your balance for as long as you like and use it whenever you need. There are no inactivity fees or expiration dates.

UMVA AI offers the cheapest AI inference for coding. Our fast models at $0.50/M input tokens are 75-90% cheaper than comparable models from providers like OpenAI, Anthropic, and others. We achieve this by optimizing our inference infrastructure and passing the savings directly to you — no markup, no middlemen.

Refunds are handled on a case-by-case basis. If you accidentally over-deposited or encounter a billing error, contact our support team and we will review your request. Because credits are consumed instantly as you use the API, used credits are generally non-refundable, but we are fair and transparent in resolving issues.

We accept Visa, Mastercard, American Express, and Verve cards, as well as cryptocurrency payments including USDT (TRC20 & BEP20), USDC, Bitcoin, and Ethereum. No KYC is required for crypto payments. We are continually adding more payment options.

UMVA AI offers four specialized models: umva-code-fast (fast and lightweight for quick edits and simple completions), umva-code-agent (fast with extended thinking for complex reasoning and multi-step tasks), umva-code-pro (advanced model for complex code generation, debugging, and analysis), and umva-code-architect (pro with extended thinking for architecture design, large-scale refactoring, and system planning). All models feature a 1 million token context window.

For quick edits, simple completions, and low-latency responses, use umva-code-fast. For multi-step reasoning, debugging, and complex analysis, use umva-code-agent. For the most sophisticated code generation, architecture design, and system-level tasks, use umva-code-pro or umva-code-architect. You can easily switch between models at any time in the IDE or API.

All UMVA AI text models feature a 1 million token (1M) context window. This means you can input up to roughly 750,000 words or 1.5 million characters in a single prompt. This allows you to process entire codebases, lengthy documentation, long conversation histories, or massive datasets without needing to chunk or truncate — leading to more accurate and context-aware responses.

All UMVA AI models support up to 384,000 (384K) output tokens per request. This is significantly higher than most providers who limit output to 4K-16K tokens. With this generous limit, you can generate entire files, comprehensive documentation, or long-form analysis in a single response without needing multiple follow-ups.

Yes. UMVA AI fully supports tool use (function calling). You can define custom functions, tools, and structured outputs, and our models will intelligently decide when and how to invoke them. This makes UMVA AI ideal for building autonomous agents, automated workflows, and complex multi-step applications.

Yes. Our models support file uploads including code files, text documents, JSON, CSV, PDF, and images in the chat interface. The AI can read and analyze the contents of uploaded files, extract data, generate summaries, and write code based on the provided materials. Native image understanding is supported for extracting text and structured information.

UMVA AI excels at code generation across all major languages and frameworks. Our models are specifically optimized for coding tasks — they understand project context, follow existing code patterns, generate idiomatic code, and handle complex algorithms, API integrations, data processing, and full-stack development. With the 1M context window, they can reason across your entire codebase for more coherent results.

Yes. umva-code-agent and umva-code-architect feature extended thinking capabilities, allowing the model to break down complex problems into step-by-step reasoning before generating a response. This produces more accurate results for debugging, algorithm design, data analysis, architecture planning, and other tasks requiring deep reasoning.

Yes. The UMVA AI API is fully compatible with the OpenAI API specification. Any client library, SDK, framework, or application that works with OpenAI's API will work with UMVA AI by simply changing the base URL and API key. This includes popular tools like LangChain, LlamaIndex, Vercel AI SDK, and the official OpenAI Python/Node.js libraries.

The UMVA AI API base URL is https://api.umva.ai/v1. The endpoint structure follows OpenAI conventions: /v1/chat/completions for chat, /v1/models for listing available models, and /v1/completions for legacy completions. Simply replace the OpenAI base URL with ours and use your UMVA API key for authentication.

Authentication is done via Bearer token in the Authorization header. Generate an API key from your UMVA dashboard under the API Keys section, then include it as follows: Authorization: Bearer YOUR_API_KEY. All communication must be over HTTPS and API keys are hashed using bcrypt for storage.

Because our API is OpenAI-compatible, you can use any OpenAI client library including the official OpenAI Python SDK, OpenAI Node.js SDK, LangChain, LlamaIndex, the Vercel AI SDK, and community libraries in Go, Rust, Java, Ruby, PHP, and Swift. Just change the base URL to https://api.umva.ai/v1.

Rate limits depend on your account tier. Standard accounts start at 100 requests per minute (RPM) and 1,000,000 tokens per minute (TPM). Higher limits are available for power users and enterprise customers. You can view your current usage and limits in the API dashboard. Contact support if you need a custom rate limit increase.

Yes. The API supports server-sent events (SSE) streaming for real-time responses. Set stream: true in your request body, and the server will progressively send tokens as they are generated. This follows the exact same SSE format as the OpenAI streaming API, making it a drop-in replacement for existing streaming integrations.

The API returns standard HTTP status codes and JSON error responses following OpenAI conventions. Common codes include 200 (success), 400 (bad request — malformed input), 401 (unauthorized — invalid API key), 429 (rate limited), and 500 (server error). Each error response includes an error object with a message and type field for programmatic handling.

The UMVA AI IDE is a free, native AI-powered code editor that runs on macOS, Windows, and Linux. It includes built-in AI chat, inline code completion, git integration, a full terminal, and syntax highlighting for hundreds of languages.

The UMVA AI IDE includes AI chat for natural language coding assistance, inline code completion as you type, code explanation, refactoring suggestions, bug detection, test generation, and intelligent code navigation. All features are powered by UMVA AI models and deduct from your wallet balance with no additional subscription fees.

Yes. UMVA AI works with VS Code through the Continue.dev extension. Simply install the extension from the VS Code marketplace, configure it with your UMVA API key and endpoint, and start using UMVA AI models for code completion, chat, and refactoring within VS Code.

Yes. UMVA AI integrates with all JetBrains IDEs (IntelliJ IDEA, PyCharm, WebStorm, GoLand, CLion, Rider, etc.) through the Continue.dev plugin available in the JetBrains Marketplace. You can also configure it as a custom OpenAI-compatible provider in JetBrains' built-in AI Assistant feature.

Yes. Neovim users can integrate UMVA AI through the Continue.dev Neovim plugin or by using the llm.nvim plugin configured with our OpenAI-compatible endpoint. You can also use direct curl commands or custom Lua scripts to call the API from within Neovim for chat and inline completions.

UMVA AI works with Cursor by configuring a custom API endpoint in Cursor's settings. Point Cursor's AI provider to https://api.umva.ai/v1 with your UMVA API key. You can then use UMVA AI models for Cursor's inline editing, AI chat, and code generation features at significantly lower cost than Cursor's built-in pricing.

Download the installer for your operating system from our website at umva.us/umva-ai. We provide native installers for macOS (.dmg), Windows (.exe), and Linux (.AppImage or .deb). Installation takes less than a minute. After launching, log in with your UMVA account and you are ready to start coding with AI.

No. We never use your prompts, code, or conversations to train or improve our AI models. Your data is processed solely to fulfill your request and is immediately discarded. We believe your code and intellectual property belong to you, not to us.

No. We do not store your prompts or generated responses on our servers. Once the response is delivered to you, the data is gone. Aggregated usage metrics (token counts, model used, timestamps) are collected for billing and operations only. No conversation content is retained.

All communication between your IDE, API client, or browser and our servers is encrypted using TLS 1.3. Data at rest is encrypted using AES-256. API keys are hashed with bcrypt before storage. We follow industry-standard security practices to protect your data at every layer.

API keys are hashed using bcrypt and never stored in plaintext. We also support key rotation via the dashboard, and you can generate, revoke, or regenerate keys at any time. All API communication requires HTTPS, and keys are transmitted only in the Authorization header, never in URLs or logs.

UMVA AI is built with privacy-first principles. We are GDPR-compliant and follow OWASP security best practices. We do not share data with third-party AI providers. If you require SOC 2, HIPAA, or other specific compliance certifications for enterprise use, please contact our sales team to discuss dedicated deployment options.

Creating an account is free and takes less than 30 seconds. Go to the UMVA website, click Sign Up, enter your email address and a password, and verify your email. No credit card, phone number, or identity documents are required. You can start using the platform immediately after verification.

No. UMVA AI does not require KYC (Know Your Customer) verification for standard accounts. You can use card or cryptocurrency payments without submitting any identity documents. Enterprise accounts may require business verification for invoicing and credit terms, but never for standard usage.

API keys are managed from your UMVA dashboard under the API Keys section. You can generate new keys, view existing keys (partial display for security), revoke compromised keys, and regenerate keys at any time. We recommend rotating keys periodically and using separate keys for development and production environments.

We offer email support at support@umva.us for billing, technical issues, and general inquiries. Response times are typically within 24 hours. Enterprise customers receive priority support with faster response times, a dedicated account manager, and optional phone or video call support.

Yes. Enterprise customers can request a dedicated deployment with isolated infrastructure, including dedicated GPU clusters, a private API endpoint, and custom rate limits. This is ideal for organizations with high-volume workloads, strict compliance requirements, or sensitive data handling needs. Contact our sales team for a custom quote.

Yes, we offer custom model fine-tuning for enterprise customers. We can fine-tune models on your proprietary codebases, documentation, or domain-specific data to create models that understand your organization's unique patterns, style guides, and internal APIs. This requires a dedicated deployment and is billed separately.

Yes. If your team exceeds 100 million tokens per month, we offer volume discounts that significantly reduce per-token pricing. The discount scales with usage — higher volume means lower rates. Contact our enterprise sales team for a customized pricing plan tailored to your team's projected usage.

Enterprise customers receive a Service Level Agreement (SLA) guaranteeing 99.9% API uptime, with financial credits for downtime. The SLA also covers maximum response latency thresholds, support response times (4 hours for critical issues), and scheduled maintenance windows with 48-hour advance notice. Dedicated uptime monitoring and monthly reports are included.