vtcode-core 0.175.5

Core library for VT Code - a Rust-based terminal coding agent
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

//! Universal LLM provider abstraction with API-specific role handling.
//!
//! The canonical trait, request/response types, and tool definitions live in
//! [`vtcode_llm::provider`]. This module re-exports them so the historical
//! `crate::llm::provider::*` import paths continue to resolve throughout
//! `vtcode-core`.

pub use vtcode_llm::provider::{
    AnthropicOptionalStringOverride, AnthropicOptionalU32Override, AnthropicRequestOverrides, AnthropicThinkingConfig,
    AnthropicThinkingDisplayOverride, AnthropicThinkingModeOverride, AssistantPhase, BorrowedLLMStream,
    ChoiceDecisionAnswer, ChoiceDecisionRequest, ChoiceDecisionResponse, CodingAgentSettings, ContentPart,
    ContextWindowProvider, DECISIONS_MODEL, DecisionChoiceOption, DecisionProbability, FallbackModel, FinishReason,
    FunctionCall, FunctionDefinition, GrammarDefinition, LLMError, LLMErrorMetadata, LLMNormalizedStream, LLMProvider,
    LLMRequest, LLMResponse, LLMStream, LLMStreamEvent, Message, MessageClearAt, MessageContent, MessageRole,
    NormalizedStreamEvent, ParallelToolConfig, PreparedResponsesRequest, PromptCacheProfile, ProviderCapabilities,
    ReasoningSource, ResponsesCompactionOptions, ResponsesContinuationState, ShellToolDefinition,
    SpecificFunctionChoice, SpecificToolChoice, ToolCall, ToolChoice, ToolDefinition, ToolNamespace,
    ToolSearchAlgorithm, Usage, get_cached_capabilities, is_standard_decisions_endpoint,
    prepare_openai_responses_request, prepare_responses_continuation_request, records_responses_continuation_state,
    responses_continuation_key, supports_responses_chaining, uses_incremental_responses_history,
};