Protocols

Configuring protocols

Tragentics supports six interoperability protocols that allow external systems and other agents to discover and communicate with your agent. Configure them from the Protocols page in the sidebar.

The Protocols page starts with the Capabilities view, which is where you shape the capability schema that feeds generated protocol cards. Then switch to the Protocols view to enable individual protocols, rate limits, and endpoint overrides. For the schema side specifically, see Capability schema.

Selecting an agent

Click Protocols in the sidebar, then select the agent you want to configure from the dropdown at the top. The page has a Capabilities view and a Protocols view. The protocols side displays six protocol cards, one for each supported standard.

The six protocol cards

Each card represents a protocol your agent can participate in:

Protocol	Description
ACP	Agent Communication Protocol — native agent manifest and capability card
MCP	Model Context Protocol — tool-based integration for LLM tool calling
OpenAI	OpenAI Responses API for multi-provider agentic workflows
A2A	Agent-to-Agent protocol for task-based agent communication
DID	Decentralized Identifier — W3C standard for verifiable digital identity
ANP	Agent Network Protocol — inter-agent networking and coordination

ACP is the native protocol — its card content is populated from your agent's Settings page, not auto-generated. The other five protocols compute their cards on-the-fly from your agent's profile data. To change what appears in the ACP card, edit your agent's settings (name, description, capabilities, formats, models, etc.).

Enabling and disabling protocols

Each protocol card has a toggle switch at the top right. Flip it on to enable the protocol for your agent. The toggle takes effect immediately — your agent appears in discovery results for that protocol as soon as enabled.

Disabling a protocol removes the agent from that protocol's discovery endpoints. For communication protocols (ACP, MCP, OpenAI, A2A, ANP), relay requests targeting the disabled protocol are also rejected. DID is discovery-only — it has no relay. Existing connections through other protocols are not affected.

You can enable multiple protocols simultaneously. An agent with MCP, A2A, and OpenAI all enabled will appear in all three discovery catalogs and accept relay calls in all three formats.

Rate limits

Communication protocol cards (ACP, MCP, OpenAI, A2A, ANP) expose two configurable rate limits — one for discovery requests and one for relay requests. DID exposes only a discovery rate limit (no relay). These control how many requests per minute external callers can make to your agent through that protocol.

Rate limits are per-agent, per-protocol, per-IP. Adjust them based on the capacity of your endpoint. The values are enforced server-side by the relay routes.

Endpoint URL overrides

By default, relay calls for communication protocols are forwarded to your agent's main endpoint URL. If you need protocol-specific routing — for example, a dedicated MCP server on a different port — set a custom URL override on that protocol's card. DID does not have an endpoint override (it is an identifier, not a communication protocol).

To set an override, enter the full URL in the endpoint field on the protocol card and save. To clear it and revert to the default endpoint, remove the URL and save. The relay route checks for a protocol-specific override first, then falls back to the main endpoint URL.

Generated card JSON preview

Each protocol card includes a JSON preview showing exactly what external systems will receive when they query the discovery endpoint. This preview updates in real time as you toggle the protocol, change rate limits, or modify capabilities.

The preview is read-only. To change what appears in the JSON, update the corresponding settings — capabilities, formats, models, rate limits — on the Settings page.

Each communication protocol card also includes a Proxy Call Example section showing how to use the X-Tragentics-Protocol header on internal proxy routes, and an Allow External Invocation toggle that controls whether unauthenticated callers can reach your agent through the relay. See Enabling external invocation for details.

Readiness checks

The Protocols tab shows a readiness indicator for each protocol. This reflects your agent's profile completeness — whether the required fields (endpoint URL, service description, capabilities) are populated. A protocol can be enabled even if readiness is incomplete, but the generated card will have missing fields.

Check the readiness indicator

Each protocol card shows which fields are populated and which are missing. Green checks indicate complete fields.

Fill in missing fields

Navigate to the Settings tab to add missing capabilities, formats, or descriptions. Return to Protocols to verify the readiness updated.

Enable the protocol

Once readiness is satisfactory, toggle the protocol on. Your agent immediately becomes discoverable through that protocol.

With protocols enabled, you can allow external systems to call your agent without Tragentics authentication. See Enabling external invocation →