# What changed in @langchain/ollama

> 14 changes in @langchain/ollama 0.3 to 1 from its official release notes, upgrade guides and advisories: 1 breaking, 1 security, 5 new APIs and options. Each one cited. Sources checked 2026-10-19. HTML: https://unstale.sh/npm/@langchain/ollama/

To see which of these apply to a repository, run `npx @unstale/cli init` in it.

## @langchain/ollama 1

### Breaking changes

- When aborted during streaming with streaming callback handlers, invoke() throws ModelAbortError containing the accumulated partialOutput; an already-aborted signal is checked immediately. (1.2.2) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.2.2)

### New APIs and options

- Structured output now supports standard schema, allowing developers to use standard schemas with Ollama. (1.2.6) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.2.6)
- The Ollama LLM class now supports the think parameter. (1.2.0) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.2.0)
- Adds support for specifying a custom baseUrl. (1.1.0) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.1.0)
- Adds support for native structured outputs in Ollama. (1.1.0) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.1.0)
- Thinking content is separated into the reasoning_content field. (1.0.3) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.0.3)

### Behavior and default changes

- Adds native streamEvents event conversion for Ollama. (1.3.0) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.3.0)
- Runnable traces include package version metadata in LangSmith trace metadata, stamped at construction time. (1.2.5) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.2.5)

### Fixes and other changes

- Preserves tool_calls when AIMessage content is a string. (1.2.3) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.2.3)
- stream() throws a regular AbortError when aborted, and streaming loops check abort signals so cancellation returns early. (1.2.2) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.2.2)
- Sets additional_kwargs.reasoning_content when using streamEvents via createAgent. (1.2.1) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.2.1)
- Reasoning tokens are now properly elevated. (1.2.0) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.2.0)
- Fix compatibility with moduleResolution set to node. (1.0.2) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/ollama%401.0.2)

## @langchain/ollama 0.3

### Security

- A fix for CVE 2025 12735 was included. (0.3.58) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/0.3.58)