Getting Started with almyty
almyty is a universal API-to-AI tool gateway. It parses API schemas (OpenAPI, GraphQL, SOAP, Protobuf), auto-generates tools, and serves them via MCP, UTCP, A2A, and Agent Skills protocols. You can also create custom tools manually using HTTP, JavaScript, GraphQL, or LLM-powered definitions.
Quick Start
1. Create an Account
Sign up at app.almyty.com (opens in a new tab) and create your first organization. Every resource in almyty belongs to an organization, enabling multi-tenant collaboration.
2. Connect an API
Navigate to APIs and click Connect API. Provide:
| Field | Description |
|---|---|
| Name | A human-readable name for the API |
| Type | openapi, graphql, soap, or protobuf |
| Base URL | The root URL of the API (e.g., https://api.example.com/v1) |
| Authentication | Optional — API key, Bearer token, Basic auth, or OAuth2 |
# Example: connecting via the API
curl -X POST https://api.almyty.com/apis \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
-d '{
"name": "My REST API",
"type": "openapi",
"baseUrl": "https://api.example.com/v1"
}'3. Import a Schema
After connecting, import the API schema to discover operations automatically. almyty supports:
- OpenAPI 3.x — JSON or YAML
- GraphQL — introspection or SDL files
- SOAP — WSDL documents
- Protobuf —
.protodefinitions
You can paste the schema content directly, provide a URL, or upload a file.
curl -X POST https://api.almyty.com/apis/{id}/import-schema \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
-d '{
"schemaUrl": "https://api.example.com/v1/openapi.json",
"generateTools": true
}'4. Generate Tools
If you set generateTools: true during import, tools are created automatically.
Otherwise, navigate to the API detail page and click Generate Tools.
Each API operation becomes one tool with:
- Automatically inferred parameters (path, query, header, body)
- JSON Schema validation
- Intelligent naming from the operation ID or summary
5. Create a Gateway
Gateways expose your tools to AI agents via different protocols.
| Protocol | Use Case |
|---|---|
| MCP | Claude, Cursor, Windsurf, and other MCP-compatible clients |
| A2A | Google's Agent-to-Agent protocol for multi-agent systems |
| UTCP | Universal Tool Call Protocol for broad compatibility |
| Skills | SKILL.md files for agent skill injection |
curl -X POST https://api.almyty.com/gateways \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
-d '{
"name": "My MCP Gateway",
"type": "mcp",
"endpoint": "/my-tools"
}'6. Assign Tools to a Gateway
After creating a gateway, assign tools to control which operations are exposed:
curl -X POST https://api.almyty.com/gateways/{gatewayId}/tools/bulk \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
-d '{ "toolIds": ["tool-1", "tool-2"] }'7. Connect Your AI Agent
Point your AI agent at the gateway endpoint:
{
"mcpServers": {
"my-tools": {
"url": "https://api.almyty.com/mcp/my-org/my-tools"
}
}
}What's Next
- Tools — Learn about the five tool types and how to create custom tools
- Agents — Build multi-step AI pipelines with the visual agent builder
- Gateways — Deep dive into each gateway protocol
- API Reference — Full REST API documentation