Skip to content

Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors

Rust

docs(llm): describe Anthropic auto tool fallback

AkitaOnRails341db8949a2b
diff --git a/crates/ai-memory-llm/src/lib.rs b/crates/ai-memory-llm/src/lib.rs
index 51d5a204..e7db3e75 100644
--- a/crates/ai-memory-llm/src/lib.rs
+++ b/crates/ai-memory-llm/src/lib.rs
@@ -11,8 +11,9 @@
//! Structured-output strategies:
//!
//! * **Anthropic**: `tools[0]` is set to a single tool whose input
-//! schema we want filled, with `tool_choice = "tool"`. The
-//! model's `tool_use` content block is the structured payload.
+//! schema we want filled. Models that accept a forced tool get
+//! `tool_choice = "tool"`; models that reject it keep the default `auto`
+//! and may return either a `tool_use` block or the JSON object as text.
//! * **OpenAI**: `response_format = { type: "json_schema", strict: true }`.
//! * **OpenAI OAuth/Codex**: ChatGPT/Codex Responses API with
//! `text.format = { type: "json_schema", strict: true }`.