You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
***Experimental multiple native tool calls per turn with guardrails**: Lets the native tool protocol run multiple tools in a single assistant turn and blocks `attempt_completion()` if any tool fails in that turn, reducing the risk of partial or incorrect completions ([#9273](https://github.com/RooCodeInc/Roo-Code/pull/9273)).
19
+
***Web-evals dashboard enhancements with dynamic tool columns and UX improvements**: Adds aggregate run statistics, per-tool success metrics, and dynamic tool usage columns so you can quickly spot failing tools or exercises and compare runs without rebuilding configs ([#9592](https://github.com/RooCodeInc/Roo-Code/pull/9592)).
20
+
***Native tools as default for specific Roo Code Cloud models**: Sets the providers’ native tool protocol as the default for `minimax/minimax-m2` and `anthropic/claude-haiku-4.5` on Roo Code Cloud to reduce configuration overhead and improve tool-calling reliability ([#9586](https://github.com/RooCodeInc/Roo-Code/pull/9586)).
21
+
22
+
## Bug Fixes
23
+
24
+
***API handler refresh on tool protocol changes**: Ensures switching API profiles that only change the tool protocol still refreshes the underlying handler and parser so tool calls always use the correct configuration ([#9599](https://github.com/RooCodeInc/Roo-Code/pull/9599)).
25
+
***Native tools file reading regression for Grok Code Fast**: Restricts the single-file read behavior to XML tools so native tool calls use the standard multi-file-aware file reader and can access the workspace as expected ([#9600](https://github.com/RooCodeInc/Roo-Code/pull/9600)).
26
+
***Roo Code Cloud embeddings revert and reliability**: Removes Roo Code Cloud as an embeddings provider to prevent `codebase_search` from appearing when it is not configured and to avoid indexing getting stuck in a standby state ([#9602](https://github.com/RooCodeInc/Roo-Code/pull/9602)).
27
+
28
+
## Provider Updates
29
+
30
+
***Bedrock Anthropic Claude Opus 4.5 for global inference**: Adds the Anthropic Claude Opus 4.5 Bedrock model to the global inference model list so it can be used automatically anywhere global inference is supported, with no extra setup ([#9595](https://github.com/RooCodeInc/Roo-Code/pull/9595)).
description: This patch release adds AWS Bedrock embeddings support, improves native tool and provider behavior, and fixes several reliability issues.
3
+
keywords:
4
+
- roo code 3.34.6
5
+
- new features
6
+
- bug fixes
7
+
image: /img/v3.34.6/v3.34.6.png
8
+
---
9
+
10
+
# Roo Code 3.34.6 Release Notes (2025-11-26)
11
+
12
+
This release adds AWS Bedrock embeddings support for code indexing, improves native tool and provider behavior, and fixes several reliability issues across Bedrock, Vertex, and native tooling.
***AWS Bedrock embeddings for code indexing**: Adds support for using AWS Bedrock embeddings in code indexing so teams that standardize on Bedrock can reuse their existing infrastructure when indexing repos for Roo-based navigation and search (PR #9475).
19
+
20
+
## QOL Improvements
21
+
22
+
***Native tool calling for Mistral**: Adds native tool calling support for the Mistral provider, enabling more reliable tool workflows and better multi-step automation when using Mistral models (PR #9625).
23
+
***Parallel native tool execution via OpenAI protocol**: Wires the `MULTIPLE_NATIVE_TOOL_CALLS` experiment to OpenAI's `parallel_tool_calls` capability so multiple tools can run in parallel under the OpenAI-compatible protocol, improving throughput for tool-heavy tasks (PR #9621).
24
+
***Fine-grained tool streaming for OpenRouter Anthropic**: Adds fine-grained tool streaming support for Anthropic models on OpenRouter so tool call responses stream more smoothly and stay aligned with Anthropic's tool semantics (PR #9629).
25
+
***Global inference selection for Bedrock with cross-region**: Allows global inference to pick Bedrock models correctly even when cross-region routing is enabled, reducing the risk of mismatched regions or unavailable models during automatic selection (PR #9616).
26
+
27
+
## Bug Fixes
28
+
29
+
***Vertex Anthropic content block filtering**: Filters out non-Anthropic content blocks before sending requests to the Vertex Anthropic API so mixed or unsupported blocks no longer cause request failures (#9583, PR #9618).
30
+
***WriteToFileTool partial content safety**: Restores a missing `content` undefined check in `WriteToFileTool.handlePartial()` to prevent rare crashes or incorrect writes when partial content is unavailable (#9611, PR #9614).
31
+
***Model cache avoids persisting empty responses**: Prevents the model cache from persisting empty API responses so configuration and model metadata do not get silently replaced with blank entries (#9597, PR #9623).
32
+
***Skip access_mcp_resource when no resources exist**: Excludes the `access_mcp_resource` tool from native tool runs when an MCP server does not expose any resources, avoiding confusing no-op calls and making tool menus cleaner (PR #9615).
33
+
***Inline terminal and indexing defaults**: Updates default settings for the inline terminal and codebase indexing to better match common workflows out of the box, reducing the amount of manual tuning needed after installation (PR #9622).
34
+
***Native new_task tool_result timing**: Defers emitting the `new_task` tool_result in the native protocol until the corresponding subtask actually finishes, eliminating premature “done” signals and keeping task state accurate for downstream tools (PR #9628).
Copy file name to clipboardExpand all lines: docs/update-notes/v3.34.mdx
+13-1Lines changed: 13 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -14,7 +14,7 @@ image: /img/v3.34.0/v3.34.0.png
14
14
15
15
# Roo Code 3.34 Release Notes (2025-11-21)
16
16
17
-
Roo Code 3.34 combines Browser Use 2.0, the Baseten provider, OpenAI-compatible improvements, and refined onboarding and native tool descriptions with patches that add Roo Code Cloud eval support, clearer image generation prompts, todo list cleanup, cloud sync fixes, Claude Opus 4.5 across Roo Code Cloud, OpenRouter, Anthropic, and Vertex, plus new provider reliability and model updates, free Black Forest Labs image generation on Roo Code Cloud, and native MCP tool reliability fixes.
17
+
Roo Code 3.34 combines Browser Use 2.0, the Baseten provider, OpenAI-compatible improvements, and refined onboarding and native tool descriptions with patches that add Roo Code Cloud eval support, clearer image generation prompts, todo list cleanup, cloud sync fixes, Claude Opus 4.5 across Roo Code Cloud, OpenRouter, Anthropic, and Vertex, AWS Bedrock embeddings for code indexing, improved native tool workflows for Mistral and OpenAI-compatible providers, smoother Anthropic tool streaming on OpenRouter, better Bedrock global inference selection, plus new provider reliability and model updates, free Black Forest Labs image generation on Roo Code Cloud, and native MCP tool reliability fixes.
@@ -36,6 +36,13 @@ Browser Use now supports a more capable "2.0" experience ([#8941](https://github
36
36
-**Clearer image generation prompts**: The full prompt and path for image generation now appear directly in the chat UI with clearer spacing and typography, making it easier to inspect, debug, and reuse prompts ([#9505](https://github.com/RooCodeInc/Roo-Code/pull/9505), [#9522](https://github.com/RooCodeInc/Roo-Code/pull/9522)).
37
37
-**Eval jobs on Roo Code Cloud**: You can now run evaluation jobs directly on Roo Code Cloud models, reusing the same managed models and job tokens as regular cloud runs ([#9492](https://github.com/RooCodeInc/Roo-Code/pull/9492), [#9522](https://github.com/RooCodeInc/Roo-Code/pull/9522)).
38
38
-**XML tool protocol stays in sync with configuration**: Tool runs that use the XML protocol now correctly track the configured tool protocol after configuration updates, preventing rare parser-state errors when switching between XML and native tools ([#9535](https://github.com/RooCodeInc/Roo-Code/pull/9535))).
39
+
-**Experimental multiple native tool calls per turn with guardrails**: Lets the native tool protocol run multiple tools in a single assistant turn and blocks `attempt_completion()` if any tool fails in that turn, reducing the risk of partial or incorrect completions ([#9273](https://github.com/RooCodeInc/Roo-Code/pull/9273)).
40
+
-**Web-evals dashboard enhancements with dynamic tool columns and UX improvements**: Adds aggregate run statistics, per-tool success metrics, and dynamic tool usage columns so you can quickly spot failing tools or exercises and compare runs without rebuilding configs ([#9592](https://github.com/RooCodeInc/Roo-Code/pull/9592)).
41
+
-**Native tools as default for specific Roo Code Cloud models**: Sets the providers’ native tool protocol as the default for `minimax/minimax-m2` and `anthropic/claude-haiku-4.5` on Roo Code Cloud to reduce configuration overhead and improve tool-calling reliability ([#9586](https://github.com/RooCodeInc/Roo-Code/pull/9586)).
42
+
-**Native tool calling for Mistral**: Adds native tool calling support for the Mistral provider, enabling more reliable tool workflows and better multi-step automation when using Mistral models ([#9625](https://github.com/RooCodeInc/Roo-Code/pull/9625)).
43
+
-**Parallel native tool execution via OpenAI protocol**: Wires the `MULTIPLE_NATIVE_TOOL_CALLS` experiment to OpenAI's `parallel_tool_calls` capability so multiple tools can run in parallel under the OpenAI-compatible protocol, improving throughput for tool-heavy tasks ([#9621](https://github.com/RooCodeInc/Roo-Code/pull/9621)).
44
+
-**Fine-grained tool streaming for OpenRouter Anthropic**: Adds fine-grained tool streaming support for Anthropic models on OpenRouter so tool call responses stream more smoothly and stay aligned with Anthropic's tool semantics ([#9629](https://github.com/RooCodeInc/Roo-Code/pull/9629)).
45
+
-**Global inference selection for Bedrock with cross-region**: Allows global inference to pick Bedrock models correctly even when cross-region routing is enabled, reducing the risk of mismatched regions or unavailable models during automatic selection ([#9616](https://github.com/RooCodeInc/Roo-Code/pull/9616)).
39
46
40
47
## Bug Fixes
41
48
@@ -49,6 +56,9 @@ Browser Use now supports a more capable "2.0" experience ([#8941](https://github
49
56
-**Skip unsupported Gemini content blocks safely**: Gemini conversations on Vertex AI now skip unsupported metadata blocks (such as certain reasoning or document types) with a warning instead of failing the entire thread, keeping long-running chats stable ([#9537](https://github.com/RooCodeInc/Roo-Code/pull/9537))).
50
57
-**Native MCP tool names preserved in history**: Native mode now keeps the real dynamic MCP tool names (such as `mcp_serverName_toolName`) in the API history instead of teaching the model a fake `use_mcp_tool` name, so follow-up calls pick the right tools and tool suggestions stay consistent ([#9559](https://github.com/RooCodeInc/Roo-Code/pull/9559)).
51
58
-**Native tools condensing keeps tool_use context**: When condensing long conversations that use native tools, required `tool_use` and `tool_result` blocks are preserved in the summary message, preventing 400 errors and avoiding lost context during follow-up turns ([#9582](https://github.com/RooCodeInc/Roo-Code/pull/9582)).
59
+
-**API handler refresh on tool protocol changes**: Ensures switching API profiles that only change the tool protocol still refreshes the underlying handler and parser so tool calls always use the correct configuration ([#9599](https://github.com/RooCodeInc/Roo-Code/pull/9599)).
60
+
-**Native tools file reading regression for Grok Code Fast**: Restricts the single-file read behavior to XML tools so native tool calls use the standard multi-file-aware file reader and can access the workspace as expected ([#9600](https://github.com/RooCodeInc/Roo-Code/pull/9600)).
61
+
-**Roo Code Cloud embeddings revert and reliability**: Removes Roo Code Cloud as an embeddings provider to prevent `codebase_search` from appearing when it is not configured and to avoid indexing getting stuck in a standby state ([#9602](https://github.com/RooCodeInc/Roo-Code/pull/9602)).
52
62
53
63
## Provider Updates
54
64
@@ -60,3 +70,5 @@ Browser Use now supports a more capable "2.0" experience ([#8941](https://github
60
70
-**Reliable LiteLLM model refresh after credential changes**: Clicking **Refresh Models** after changing your LiteLLM API key or base URL now immediately reloads the model list using the new credentials, so you do not need to clear caches or restart VS Code, while background refreshes still benefit from caching for speed ([#9536](https://github.com/RooCodeInc/Roo-Code/pull/9536))).
61
71
-**Black Forest Labs image generation on Roo Code Cloud**: Use the free `bfl/flux-2-pro:free` model on Roo Code Cloud for high-quality image generation without unexpected charges, powered by the `images_api` image generation method for compatible models ([#9587](https://github.com/RooCodeInc/Roo-Code/pull/9587)).
62
72
-**Black Forest Labs models on OpenRouter**: Adds Black Forest Labs FLUX.2 Flex and FLUX.2 Pro image generation models via OpenRouter, giving you additional high-quality options when you prefer to use your OpenRouter account for image generation ([#9589](https://github.com/RooCodeInc/Roo-Code/pull/9589)).
73
+
-**Bedrock Anthropic Claude Opus 4.5 for global inference**: Adds the Anthropic Claude Opus 4.5 Bedrock model to the global inference model list so it can be used automatically anywhere global inference is supported, with no extra setup ([#9595](https://github.com/RooCodeInc/Roo-Code/pull/9595)).
74
+
-**AWS Bedrock embeddings for code indexing**: Adds support for using AWS Bedrock embeddings in code indexing so teams that standardize on Bedrock can reuse their existing infrastructure when indexing repos for Roo-based navigation and search ([#9475](https://github.com/RooCodeInc/Roo-Code/pull/9475)).
0 commit comments