# Apertis AI — Full Service Description

> AI control plane for coding agents and AI apps: OpenAI-compatible routing, budgets, model policy, prompt cache, and failover.

## About

Apertis AI is an AI control plane for teams that run coding agents and AI applications. It provides one OpenAI-compatible gateway for routing requests across 500+ models and 30+ providers while adding operational controls such as virtual keys, budgets, model policy, routing health, prompt cache, and automatic failover.

The primary brand position is control and reliability for agent traffic. Model breadth is supporting proof: Apertis keeps access broad while giving teams a central place to govern spend, provider choice, and rollout behavior.

## Who Apertis Serves

- Engineering teams standardizing access for coding agents and AI tools
- Platform teams that need centralized model policy and budget controls
- Developers migrating OpenAI-compatible applications to a multi-provider gateway
- Teams comparing provider cost, context windows, routing health, and failover options
- Organizations that want a managed alternative to self-hosted routing infrastructure

## Control-Plane Capabilities

### Virtual Keys

Teams can issue controlled API keys for applications, environments, users, and coding agents. This gives operators a way to separate usage, rotate access, and connect spend to the right workflow.

### Budgets And Usage Controls

Apertis supports budget-aware rollout for AI applications and coding tools. Teams can track usage, compare costs, and prevent uncontrolled agent spend from spreading across unmanaged provider accounts.

### Model Policy

Model policy helps teams decide which models and providers are available for a workflow. It supports governed access without forcing application code to understand every upstream provider.

### Routing Health And Failover

Requests can route through available provider channels, with automatic failover when a primary channel is unavailable. This improves reliability for coding agents and production AI apps.

### Prompt Cache

Prompt cache reduces cost and latency for repeated prompts or reusable prefixes. Cache reads are free, and teams can use the feature without changing their client integration.

### Model And Provider Coverage

Apertis provides routed access to 500+ models from 30+ providers, including OpenAI, Anthropic, Google, AWS Bedrock, Azure OpenAI, Mistral, Cohere, DeepSeek, xAI, Alibaba/Qwen, Baidu, Tencent, Zhipu, Cloudflare Workers AI, and more.

## API Endpoints

All API endpoints require an Apertis API key sent as a Bearer token.

### Base URL

```text
https://api.apertis.ai/v1
```

### Text Generation

- `POST /v1/chat/completions` — OpenAI-compatible chat completions for frontier, open, and provider-hosted models.
- `POST /v1/completions` — Legacy text completions for compatible models.
- `POST /v1/messages` — Native Anthropic Messages API format for Anthropic-style clients.

### Other Modalities

- `POST /v1/embeddings` — Text embedding generation.
- `POST /v1/images/generations` — Image generation through supported image models.
- `POST /v1/audio/speech` — Text-to-speech.
- `POST /v1/audio/transcriptions` — Speech-to-text.
- `POST /v1/audio/translations` — Audio translation.
- `POST /v1/moderations` — Content moderation through supported moderation models.

### Models

- `GET /v1/models` — List available models.
- `GET /v1/models/:model` — Get model details.

## Integration Pattern

Apertis is compatible with the OpenAI SDK pattern. Existing applications usually change the API key and base URL while keeping the same request shape.

```python
from openai import OpenAI

client = OpenAI(
    api_key="your-apertis-key",
    base_url="https://api.apertis.ai/v1",
)

response = client.chat.completions.create(
    model="claude-sonnet-4-20250514",
    messages=[{"role": "user", "content": "Hello"}],
)
```

## Web Search Suffix

When enabled for an account and model, adding `:web` to a model ID can trigger web-search augmentation:

```python
response = client.chat.completions.create(
    model="gpt-4o-mini:web",
    messages=[{"role": "user", "content": "Summarize today's AI infra news"}],
)
```

## Pricing And Plans

- Pay-as-you-go: no monthly fee; pay based on usage.
- Coding Plan Lite: monthly plan for individual developer workflows.
- Coding Plan Pro: higher quota plan for power users and frequent coding-agent usage.
- Coding Plan Max: larger plan for teams and heavier agent workloads.
- Enterprise pilot: managed Team Gateway rollout for organizations that need central controls and support.

All plans can use the Apertis gateway capabilities available to the account, including model routing, prompt cache, and failover.

## SDK And Tool Integrations

- OpenAI Python SDK: compatible through base URL and API key changes.
- OpenAI JavaScript SDK: compatible through base URL and API key changes.
- Vercel AI SDK: Apertis provider support.
- LiteLLM: supported as a proxy target.
- LlamaIndex: supported integration.
- Kilo Code: supported provider workflow.
- MCP: Apertis MCP server metadata is available for agent clients.

## Compatible Coding Tools

- Claude Code
- Cursor
- OpenCode
- Aider
- Cline
- Continue
- Kilo Code

## Authoritative Discovery Files

- [llms.txt](https://apertis.ai/llms.txt): concise AI-facing site summary.
- [OpenAPI Description](https://apertis.ai/docs/openapi.yaml): machine-readable API description.
- [API Catalog](https://apertis.ai/.well-known/api-catalog): API discovery metadata.
- [Agent Skills Index](https://apertis.ai/.well-known/agent-skills/index.json): agent-facing skills index.
- [MCP Server Card](https://apertis.ai/.well-known/mcp/server-card.json): MCP server metadata.
- [Sitemap](https://apertis.ai/sitemap.xml): canonical public URL map.
- [Robots](https://apertis.ai/robots.txt): crawler and content-signal policy.

## Core Pages

- [Homepage](https://apertis.ai/): control-plane positioning, hero overview, supported providers, and pilot CTA.
- [Models](https://apertis.ai/models): searchable model catalog with pricing, providers, context windows, and feature metadata.
- [Pricing](https://apertis.ai/subscribe): Coding Plan and pay-as-you-go information.
- [Enterprise Pilot](https://apertis.ai/enterprise/form): contact form for Team Gateway pilots.
- [Trust Center](https://apertis.ai/trust): security controls, data handling, compliance posture, and subprocessors.
- [Changelog](https://apertis.ai/changelog): product updates, model additions, and release notes.
- [Cost Calculator](https://apertis.ai/calculator): usage and cost comparison tool.
- [Model Comparison](https://apertis.ai/models/compare): side-by-side model comparison.

## Research And Comparison Pages

- [What Is an AI API Gateway](https://apertis.ai/research/what-is-ai-api-gateway): explains the AI gateway pattern.
- [Reduce AI API Costs](https://apertis.ai/research/reduce-ai-api-costs): guide to cost reduction through routing, caching, and model choice.
- [Setup Claude Code with Apertis](https://apertis.ai/research/setup-claude-code-with-apertis): coding-agent setup guide.
- [Apertis vs OpenRouter](https://apertis.ai/compare/openrouter): comparison of managed routing, model coverage, and workflow fit.
- [Apertis vs Together AI](https://apertis.ai/compare/together-ai): comparison with Together AI.
- [Apertis vs LiteLLM](https://apertis.ai/compare/litellm): comparison with self-hosted proxy infrastructure.

## Trust, Privacy, And Legal

- Apertis does not train models on user data.
- Requests are routed to upstream providers according to provider and account configuration.
- TLS is used for data in transit.
- API keys are protected in storage.
- See the Trust Center, Privacy Policy, Terms of Service, and Refund Policy for current legal and security details.

## Contact

- Website: https://apertis.ai
- Email: hi@apertis.ai
- Status: https://status.apertis.ai
- GitHub: https://github.com/apertis-ai
- API docs: https://docs.apertis.ai
- Terms: https://apertis.ai/terms
- Privacy: https://apertis.ai/privacy