Module agentforge4j.llm.api

Package com.agentforge4j.llm.api

package com.agentforge4j.llm.api

Public LLM invocation contract shared by runtime callers and provider adapters.

Key types:

• LlmClient — executes requests for one provider id
• LlmExecutionRequest — immutable invocation parameters
• LlmExecutionResponse — model output and optional token usage
• LlmInvocationException — failures from clients or transport
• LlmRetryPolicy — optional retry settings exposed by clients
• TokenUsageReport — provider token usage for an invocation
• PromptLayerBoundaries — stable-prefix layer byte boundaries

Related Packages

Module	Package	Description
agentforge4j.llm	com.agentforge4j.llm	LLM implementation support: client resolution, service discovery, and provider wiring.
agentforge4j.llm.fake	com.agentforge4j.llm.fake

Class	Description
LlmClient	Executes LLM requests for a single registered provider (for example OpenAI, Ollama, or Claude).
LlmExecutionRequest	Immutable parameters for a single LLM invocation.
LlmExecutionResponse	Result of a single LlmClient invocation.
LlmInvocationException	Thrown when an LLM request fails due to network issues, HTTP errors, invalid responses, or provider-specific validation failures.
LlmInvocationIdentity	Optional run/workflow/step/agent identity for a single LLM invocation.
LlmRetryPolicy	Immutable retry settings for a retrying LLM client: attempt cap, decorrelated jitter bounds, and optional total elapsed budget.
ModelTier	Capability tier an agent or step declares instead of a concrete, versioned model string.
ModelTierResolutionException	Thrown when a declared ModelTier cannot be resolved to a concrete model string for the selected provider — that is, no mapping exists after shipped defaults and any operator overrides have been applied.
ModelTierResolver	Resolves a declared ModelTier to a concrete, versioned model string for a given provider.
PromptLayerBoundaries	Byte offsets marking the end of each prompt layer in an assembled provider request body.
TokenUsageReport	Token usage reported by an LLM provider for a single invocation.