# Vivgrid
> Vivgrid is the Managed Skills platform: serverless LLM function calling that lets enterprise AI agents run their tools in the cloud — with observability, evaluation, and globally distributed inference built in.
## About Vivgrid
- Category: Managed Skills — cloud-hosted tools (LLM function calling) for AI agents. Personal agents can run tools on laptops; enterprise agents cannot. Vivgrid hosts skills server-side with centralized security, auditing, and instant iteration.
- Architecture: system prompts, models, and tools are managed server-side, decoupled from the client, so agents improve in minutes instead of release cycles.
- Products: Managed Skills (serverless LLM function calling), unified Model API (GPT, Claude, Gemini, DeepSeek families), agent evaluation, observability and insights, geo-distributed low-latency inference.
- Compared to agent observability tools (LangSmith, Braintrust): Vivgrid also hosts and runs agent skills in production, not only traces and evaluates them.
- Compared to model routers (OpenRouter): Vivgrid manages the full server-side agent runtime — prompts, models, and tools — not only model access.
- Company: Vivgrid is a product of Allegro US, LLC, San Francisco, California, United States. Contact: hi@vivgrid.com.
- Open source: built by the team behind the Yomo framework (https://github.com/yomorun/yomo). GitHub: https://github.com/vivgrid.
Each documentation link below points to a clean Markdown (.md) version of the page, optimized for LLMs and AI agents.
## Get Started

- [Alchemist - Agent Evaluation Tool](https://vivgrid.com/docs/agent-evaluation.md): Alchemist is Vivgrid's agent evaluation tool that enables developers to run recursive and regression evaluations to validate AI agent behavior before release.
- [Introduction](https://vivgrid.com/docs/introduction.md): Vivgrid is the Managed Skills platform — cloud-hosted LLM function calling for enterprise AI agents, with built-in observability and evaluation.
- [Managed Skills](https://vivgrid.com/docs/managed-skills.md): Build a strongly typed AI skill (Function Tool or MCP Tool) and deploy it to Vivgrid's global network in minutes with Yomo and the viv CLI.
- [Models](https://vivgrid.com/docs/models.md): All AI models supported by Vivgrid — including GPT-5, Claude Opus/Sonnet, Gemini, and DeepSeek variants for coding, text, speech, and multimodal workloads.
- [Observability & Insights](https://vivgrid.com/docs/observability.md): Vivgrid's observability features provide deep insights into AI agent performance, enabling developers to monitor, analyze, and optimize their agents effectively.
- [Quick Start](https://vivgrid.com/docs/quick-start.md): Go from your first API request to a custom AI skill running on Vivgrid's global network — in minutes.
- [Vivgrid Open Source & Startup Sponsorship](https://vivgrid.com/docs/sponsorship.md): Build, test, and scale with Vivgrid’s platform plus sponsored access to gpt-5, gemini-3, and DeepSeek-v3.
- [viv CLI — Deploy AI Skills to Vivgrid](https://vivgrid.com/docs/viv.md): Deploy AI skills to Vivgrid's global network with the viv CLI: commands for deploy, upload, create, status, logs, and vivgrid.yml configuration.

## Models

- [Models](https://vivgrid.com/docs/models.md): All AI models supported by Vivgrid — including GPT-5, Claude Opus/Sonnet, Gemini, and DeepSeek variants for coding, text, speech, and multimodal workloads.
- [claude-haiku-4-5 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/claude-haiku-4-5.md): claude-haiku-4-5 on Vivgrid: Anthropic's fast, low-cost Haiku model with a 200K context window, Messages API, and geo-distributed acceleration.
- [claude-opus-4-6 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/claude-opus-4-6.md): claude-opus-4-6 on Vivgrid: Anthropic's high-end Opus coding model with a 1M-token context window, Messages API, and geo-distributed acceleration.
- [claude-opus-4-7 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/claude-opus-4-7.md): Run Anthropic's claude-opus-4-7 on Vivgrid: a frontier coding model with a 1M-token context window, the Messages API, and geo-distributed acceleration.
- [claude-opus-4-8 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/claude-opus-4-8.md): Run Anthropic's claude-opus-4-8 on Vivgrid: the frontier Opus coding model with a 1M-token context window, the Messages API, and geo-distributed acceleration.
- [claude-opus-5 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/claude-opus-5.md): Run Anthropic's claude-opus-5 on Vivgrid: the flagship Opus coding model with a 1M-token context window, 128K max output, and worldwide geo-distributed acceleration.
- [claude-sonnet-4-6 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/claude-sonnet-4-6.md): claude-sonnet-4-6 on Vivgrid: Anthropic's balanced Sonnet model with a 1M-token context window, Messages API, and geo-distributed acceleration.
- [claude-sonnet-5 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/claude-sonnet-5.md): claude-sonnet-5 on Vivgrid: Anthropic's latest Sonnet model with a 1M-token context window, Messages API, and geo-distributed acceleration.
- [deepseek-r1-0528 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/deepseek-r1-0528.md): deepseek-r1-0528 on Vivgrid: DeepSeek's reasoning-focused R1 model with a 164K context window, available through one unified API.
- [deepseek-v3.1 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/deepseek-v3.1.md): deepseek-v3.1 on Vivgrid: DeepSeek's reliable general-purpose model with a 131K context window and low token pricing.
- [deepseek-v3.2-exp — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/deepseek-v3.2-exp.md): deepseek-v3.2-exp on Vivgrid: DeepSeek's experimental V3.2 model with a 164K context window, available through one unified API.
- [deepseek-v3.2 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/deepseek-v3.2.md): deepseek-v3.2 on Vivgrid: DeepSeek's efficient general-purpose model with a 128K context window and very low token pricing.
- [deepseek-v4-flash — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/deepseek-v4-flash.md): deepseek-v4-flash on Vivgrid: DeepSeek's fast, ultra-affordable model with a 1M-token context window and up to 384K output tokens.
- [deepseek-v4-pro — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/deepseek-v4-pro.md): deepseek-v4-pro on Vivgrid: DeepSeek's flagship coding model with a 1M-token context window, up to 384K output tokens, and competitive pricing.
- [gemini-2.0-flash — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gemini-2.0-flash.md): gemini-2.0-flash on Vivgrid: Google's efficient multimodal model with a 1M-token context window across text, image, video, audio and PDF.
- [gemini-2.5-flash — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gemini-2.5-flash.md): gemini-2.5-flash on Vivgrid: Google's fast, affordable multimodal model with a 1M-token context window across text, image, video, audio and PDF.
- [gemini-2.5-pro — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gemini-2.5-pro.md): gemini-2.5-pro on Vivgrid: Google's proven multimodal model with a 1M-token context window across text, image, video, audio and PDF.
- [gemini-3-flash-preview — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gemini-3-flash-preview.md): gemini-3-flash-preview on Vivgrid: Google's fast multimodal Gemini 3 model with a 1M-token context window and very low cached pricing.
- [gemini-3-pro-image — API, Pricing & Image Tokens | Vivgrid](https://vivgrid.com/docs/models/gemini-3-pro-image.md): gemini-3-pro-image (Nano Banana Pro) on Vivgrid: Google's highest-quality image generation and editing model, with per-image token pricing.
- [gemini-3-pro-preview — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gemini-3-pro-preview.md): gemini-3-pro-preview on Vivgrid: Google's high-end multimodal model with a 1M-token context window across text, image, video, audio and PDF.
- [gemini-3.1-flash-image — API, Pricing & Image Tokens | Vivgrid](https://vivgrid.com/docs/models/gemini-3.1-flash-image.md): gemini-3.1-flash-image (Nano Banana 2) on Vivgrid: Google's high-volume image generation and editing model, with per-image token pricing.
- [gemini-3.1-flash-lite-preview — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gemini-3.1-flash-lite-preview.md): gemini-3.1-flash-lite-preview on Vivgrid: Google's lightweight multimodal model with a ~1M-token context window at a very low price.
- [gemini-3.1-pro-preview — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gemini-3.1-pro-preview.md): gemini-3.1-pro-preview on Vivgrid: Google's high-end multimodal coding model with a ~1M-token context window across text, image, video, audio and PDF.
- [gemini-3.5-flash — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gemini-3.5-flash.md): gemini-3.5-flash on Vivgrid: Google's fast multimodal model with a 1M-token context window across text, image, video, audio and PDF.
- [gemini-3.6-flash — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gemini-3.6-flash.md): gemini-3.6-flash on Vivgrid: Google's fast agentic model with a 1M-token context window, tuned for coding loops and multi-step tool use.
- [gemini-3.7-flash — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gemini-3.7-flash.md): gemini-3.7-flash on Vivgrid: Google's latest fast agentic model with a 1M-token context window, at introductory pricing through 2026.
- [glm-5.1 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/glm-5.1.md): glm-5.1 on Vivgrid: Zhipu AI's capable coding model with a 200K context window, function calling, and competitive pricing.
- [glm-5.2 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/glm-5.2.md): glm-5.2 on Vivgrid: Zhipu AI's coding model with a 1M-token context window, function calling, and competitive pricing.
- [glm-5.3 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/glm-5.3.md): glm-5.3 on Vivgrid: Zhipu AI's coding and long-horizon agent model with a 1M-token context window, function calling, and competitive pricing.
- [gpt-4.1 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-4.1.md): gpt-4.1 on Vivgrid: OpenAI's GPT-4.1 with a ~1M-token context window and image input, available through a single unified API.
- [gpt-4o — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-4o.md): gpt-4o on Vivgrid: OpenAI's multimodal GPT-4o with a 128K context window and image input, accessible through one unified API key.
- [gpt-5-chat — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5-chat.md): gpt-5-chat on Vivgrid: a conversation-tuned GPT-5 model with a 128K context window and image input, accessed through one unified API.
- [gpt-5-mini — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5-mini.md): gpt-5-mini on Vivgrid: a fast, affordable GPT-5 model with a 272K context window, function calling, and three-region acceleration.
- [gpt-5.1-codex-max — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.1-codex-max.md): gpt-5.1-codex-max on Vivgrid: an extended-effort Codex coding model on the Responses API with a 400K context window and geo-distributed acceleration.
- [gpt-5.1-codex — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.1-codex.md): gpt-5.1-codex on Vivgrid: a Codex-tuned coding model on the Responses API with a 400K context window and three-region geo-distributed acceleration.
- [gpt-5.1 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.1.md): gpt-5.1 on Vivgrid: a general-purpose GPT-5 model with a 272K context window, text, image and audio input, and regional acceleration.
- [gpt-5.2-codex — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.2-codex.md): gpt-5.2-codex on Vivgrid: a Codex-tuned coding model on the Responses API with a 400K context window, function calling, and regional acceleration.
- [gpt-5.2 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.2.md): gpt-5.2 on Vivgrid: a capable general-purpose GPT-5 model with a 400K context window, function calling, and geo-distributed acceleration.
- [gpt-5.3-codex — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.3-codex.md): gpt-5.3-codex on Vivgrid: OpenAI's Codex-tuned coding model on the Responses API, with a 400K context window and geo-distributed acceleration.
- [gpt-5.4-mini — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.4-mini.md): gpt-5.4-mini on Vivgrid: a fast, low-cost GPT-5.4 model with a 400K context window, PDF and image input, and three-region acceleration.
- [gpt-5.4-nano — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.4-nano.md): gpt-5.4-nano on Vivgrid: the smallest, most economical GPT-5.4 model with a 400K context window and image and PDF input for high-scale tasks.
- [gpt-5.4 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.4.md): Use OpenAI's gpt-5.4 on Vivgrid: a 1.05M-token coding model with broad three-region acceleration (AMER, EMEA, APAC) and a unified, OpenAI-compatible API.
- [gpt-5.5 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.5.md): Run OpenAI's gpt-5.5 on Vivgrid: a 1.05M-token flagship coding model for agents and CLIs, with geo-distributed acceleration and a unified API.
- [gpt-5.6-luna — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.6-luna.md): gpt-5.6-luna on Vivgrid: OpenAI's fast, low-cost gpt-5.6 model with a 1.05M-token context window for high-volume agent workloads.
- [gpt-5.6-sol — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.6-sol.md): Run OpenAI's gpt-5.6-sol on Vivgrid: the flagship gpt-5.6 coding model with a 1.05M-token context window, geo-distributed acceleration, and a unified API.
- [gpt-5.6-terra — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.6-terra.md): gpt-5.6-terra on Vivgrid: OpenAI's balanced gpt-5.6 model with a 1.05M-token context window, strong coding performance, and geo-distributed acceleration.
- [gpt-5 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/gpt-5.md): gpt-5 on Vivgrid: OpenAI's GPT-5 model with a 272K context window, function calling, and three-region geo-distributed acceleration.
- [grok-4-1-fast-non-reasoning — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/grok-4-1-fast-non-reasoning.md): grok-4-1-fast-non-reasoning on Vivgrid: xAI's low-latency model with a 2M-token context window and image input, via one unified API.
- [grok-4-1-fast-reasoning — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/grok-4-1-fast-reasoning.md): grok-4-1-fast-reasoning on Vivgrid: xAI's fast reasoning model with a 2M-token context window and image input, via one unified API.
- [kimi-k2.6 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/kimi-k2.6.md): kimi-k2.6 on Vivgrid: Moonshot AI's coding model with a 262K context window, up to 256K output tokens, and function calling.
- [kimi-k3 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/kimi-k3.md): Run Moonshot AI's kimi-k3 on Vivgrid: a frontier coding model with a 1M-token context window, 128K max output, tool calling, and an OpenAI-compatible API.
- [llama3.1-8b — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/llama3.1-8b.md): llama3.1-8b on Vivgrid: Meta's efficient 8B open model with a 128K context window, available through one unified API.
- [minimax-m2.7 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/minimax-m2.7.md): minimax-m2.7 on Vivgrid: MiniMax's affordable coding model with a 205K context window, 131K max output, and APAC acceleration.
- [minimax-m3 — API, Pricing & Context Window | Vivgrid](https://vivgrid.com/docs/models/minimax-m3.md): minimax-m3 on Vivgrid: MiniMax's efficient coding model with a 512K context window, 128K max output, and very competitive pricing.

## Tutorials

- [Claude Code × Vivgrid](https://vivgrid.com/docs/tutorials/claude-code.md): How to use Vivgrid with claude-opus-4.8, claude-sonnet-4.6 in Claude Code
- [OpenClaw × Vivgrid](https://vivgrid.com/docs/tutorials/clawdbot.md): How to configure OpenClaw to use advanced LLMs through Vivgrid's unified model gateway
- [ChatGPT/Codex × Vivgrid](https://vivgrid.com/docs/tutorials/codex.md): How to use Vivgrid with gpt-5.3-codex, gpt-5.5 and gpt-5.6 in Codex
- [Elia x Vivgrid](https://vivgrid.com/docs/tutorials/elia.md): How to integrate Vivgrid to Elia
- [OpenCode × Vivgrid](https://vivgrid.com/docs/tutorials/opencode.md): How to use Vivgrid with gpt-5.5 in OpenCode.ai

## API Reference

- [Chat Completions](https://vivgrid.com/docs/api/agent/completions.md): OpenAI-compatible Chat Completions endpoint.
- [OCR](https://vivgrid.com/docs/api/model-api/ocr.md): Extracts text from an image via chat-style message with image_url content.
- [Embeddings](https://vivgrid.com/docs/api/model-api/embedding.md): Generates an embedding vector representing the input text.
- [Vibe Coding](https://vivgrid.com/docs/api/model-api/vibe-coding.md): Creates a model response. Provide text or image inputs to generate text or JSON outputs. Supports tool calling and conversation state.
- [Audio-to-Text](https://vivgrid.com/docs/api/model-api/whisper.md): Transcribes audio into the input language.
- [Text-to-Speech](https://vivgrid.com/docs/api/model-api/tts.md): Generates audio from text input (text-to-speech).
- [Image](https://vivgrid.com/docs/api/model-api/images.md): Creates an image from a text prompt (OpenAI-compatible).
- [Rerank](https://vivgrid.com/docs/api/model-api/rerank.md): Re-ranks a list of documents against a query using an embedding/rerank model.

## Marketplace

- [Autosend](https://vivgrid.com/docs/marketplace/autosend.md): Send transactional and automated emails via Autosend on Vivgrid

## Blog

- [Decouple the Agent: Why Prompts, Tools, and Models Don't Belong in Your Client](https://vivgrid.com/blog/decoupling-prompts-tools-models-from-agent-client.md): A best-practices guide for enterprise AI agents: treat system prompts as runtime config, run tools as Managed Skills, and make models a platform decision — so you can debug faster, ship AI improvements without app releases, and fix production behavior from the console in minutes.
- [Vivgrid Skills Marketplace - Autosend](https://vivgrid.com/blog/marketplace-autosend.md): Automate customer emails from your AI agents using Autosend MCP on Vivgrid. Generate and send transactional emails, notifications, and more — no code needed.
- [Why Enterprise Agents Can't Run Tools on Laptops](https://vivgrid.com/blog/why-enterprise-agents-cant-run-tools-on-laptops.md): Personal agents run tools locally. Enterprise agents can't — credentials, auditing, consistency, and iteration speed all break at company scale. The answer is Managed Skills: agent tools that run in the cloud.

## Company

- [Pricing](https://vivgrid.com/pricing.md): Free tier with pay-as-you-go pricing and a $25/mo Growth plan; $200/mo credits available to approved open-source maintainers on request.