> ## Documentation Index
> Fetch the complete documentation index at: https://aisa.one/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Supported AIsa Capabilities

> Evaluate the main AIsa capability surfaces and find the live source of truth for models, APIs, Skills, actions, discovery, and billing.

AIsa supports multiple capability surfaces. Availability is not represented by one static count: model routes, API providers, Skills, MCP descriptors, and authenticated actions have different release and status mechanisms.

Use this page to identify the correct source of truth before building against a capability.

## AI model inference

AIsa exposes model inference through `https://api.aisa.one/v1` with OpenAI-compatible and selected provider-compatible routes.

Check:

* [Live model catalog](https://aisa.one/models) for current availability.
* [Models guide](/docs/guides/models) for IDs, capability vocabulary, and endpoint guidance.
* [AI model pricing](/docs/guides/pricing/ai-model-pricing-llm-inference) for pricing guidance.

Do not infer model availability or modality support from the provider family name alone.

## Direct data and capability APIs

AIsa exposes non-chat APIs under `https://api.aisa.one/apis/v1`.

The [API Reference](/docs/api-reference) is the technical entry point for exact methods, paths, parameters, authentication, and examples. Relevant domains include search, research, financial data, social platforms, prediction markets, media, company and people data, and other providers represented in the current reference.

A page in the API reference documents a route; production use should still account for current health, pricing, rate limits, and provider-specific requirements.

## Agent Skills

[Agent Skills](/docs/agent-skills) are reusable instruction bundles for supported agent runtimes. They may combine models, APIs, scripts, safety notes, and output guidance.

Check:

* The Skill page for its intended trigger and tools.
* The Skill source repository for the current implementation.
* Configuration requirements before execution.

A Skill is not the same as an independently hosted service. It teaches an agent how to use the underlying capabilities.

## Authenticated actions

Some APIs can create, update, send, publish, follow, or otherwise change an external system. These operations may require an OAuth-connected account or additional authorization.

Treat them separately from read-only APIs:

* Verify required permissions.
* Identify the external side effect.
* Obtain confirmation when required.
* Verify the external result after execution.

## Machine-readable discovery

AIsa publishes:

* [Agent Card](https://aisa.one/.well-known/agent-card.json) for high-level A2A-style discovery.
* [MCP manifest](https://aisa.one/.well-known/mcp.json) for MCP-oriented descriptors.
* [OpenAPI specification](https://aisa.one/openapi.yaml) for request and response contracts.
* [Technical llms.txt](https://aisa.one/docs/llms.txt) for documentation routing.

Check status fields in discovery resources. A planned MCP descriptor should not be treated as a live MCP server.

## Billing and payments

Model and API usage is usage-based. Review:

* [Pricing and Billing](/docs/guides/pricing)
* [Wallet and Payments](/docs/guides/pricing/wallet)
* [Usage Logs](/docs/guides/dashboard/usage-logs)
* [Machine payments for agents](/docs/concepts/machine-payments-for-agents)

Do not assume every capability supports the same authentication or payment path.

## Verification checklist

Before relying on a capability:

1. Confirm its live model, API, Skill, or discovery entry.
2. Read the exact authentication requirements.
3. Check pricing and rate limits.
4. Test the smallest safe request.
5. Verify response shape and failure behavior.
6. For writes or payments, verify authorization and side effects separately.

For product-fit guidance, see [When to use AIsa](/docs/evaluate/when-to-use-aisa).