Open Source Dashboard - Giulio Leone

langchain-ai/langchain130k

fix(openai): add type: message to Responses API input items

Python9 giorni fa

745

langchain-ai/langchain130k

fix(langchain): normalize custom detector output to prevent KeyError in hash/mask strategies

Python10 giorni fa

711

vllm-project/vllm74k

fix: resolve chat template names before kwargs detection

Python6 giorni fa

591

docling-project/docling56k

fix(readingorder): assign FURNITURE content_layer to footer/header in container groups

Python18 giorni fa

131

mem0ai/mem050k

fix: add missing 'json' keyword to graph memory prompts (fixes #4248)

Python11 giorni fa

6691

crewAIInc/crewAI46k

fix(bedrock): group parallel tool results in single user message

Python10 giorni fa

25038

crewAIInc/crewAI46k

fix(telemetry): skip signal handler registration in non-main threads

Python18 giorni fa

44

gradio-app/gradio42k

preserve file extension when filename stem is stripped entirely in gr.File

Python11 giorni fa

15

gradio-app/gradio42k

perf: use deque for SSE pending message queues in gradio_client

Python19 giorni fa

146

gradio-app/gradio42k

fix(events): correct type hints for gr.on triggers parameter

Python10 giorni fa

123

deepspeedai/DeepSpeed42k

perf: use deque for FIFO queues in sequence parallel, superoffload, and compile

Python19 giorni fa

98

BerriAI/litellm40k

fix: exclude gpt-5.2-chat from temperature passthrough

Python19 giorni fa

1,061179

BerriAI/litellm40k

fix(anthropic): handle OAuth tokens in count_tokens endpoint

Python19 giorni fa

5,666854

BerriAI/litellm40k

fix(azure): forward realtime_protocol from config and relax api_version check for GA path

Python19 giorni fa

5,692855

BerriAI/litellm40k

fix(tools): gracefully repair truncated JSON in tool call arguments

Python17 giorni fa

23712

BerriAI/litellm40k

fix: WebSearch interception fails with thinking enabled + SpendLog dedup

Python15 giorni fa

51722

BerriAI/litellm40k

fix(streaming): prevent Vertex AI Claude content truncation when finish_reason races content

Python15 giorni fa

991

BerriAI/litellm40k

fix(bedrock): strip output_config from Bedrock Invoke requests

Python11 giorni fa

108

BerriAI/litellm40k

fix(cost_calc): update custom_llm_provider when base_model has different provider prefix

Python13 giorni fa

913

BerriAI/litellm40k

feat(anthropic): support top-level cache_control for automatic prompt caching

Python18 giorni fa

908

BerriAI/litellm40k

fix(featherless_ai): use correct FEATHERLESS_AI_API_KEY env var name

Python17 giorni fa

453

BerriAI/litellm40k

fix(models): set gpt-5.4-pro mode to responses — fixes #23014

Python9 giorni fa

3016

BerriAI/litellm40k

fix(anthropic): populate output_config when reasoning_effort is used on Claude 4.6

Python18 giorni fa

265

agno-agi/agno39k

fix: wrap input as list for non-default Azure models

Python18 giorni fa

1866

agno-agi/agno39k

fix: prevent deadlock in parallel MCP tool calls with header_provider (fixes #6094)

Python17 giorni fa

388125

HKUDS/LightRAG30k

perf: use deque for BFS queue in get_knowledge_subgraph()

Python19 giorni fa

32

CopilotKit/CopilotKit30k

fix: guard against non-object tool arguments from LLMs

TypeScript19 giorni fa

49015

CopilotKit/CopilotKit30k

fix(react): replace static KaTeX CSS import with dynamic runtime injection

TypeScript19 giorni fa

281

microsoft/semantic-kernel27k

Python: fix(python/google): preserve thought_signature in Gemini function call parts

C#18 giorni fa

48016

mlflow/mlflow25k

Add authentication support to OTLP exporter headers

Python18 giorni fa

727

vercel-labs/agent-browser23k

fix(#469): deduplicate cursor-interactive elements in snapshot -C

Rustcirca un mese fa

152

vercel-labs/agent-browser23k

fix: correct misleading output for cookies clear and tab close

Rust19 giorni fa

124

modelcontextprotocol/python-sdk22k

fix(oauth): include client_id in token request body for client_secret_post

Python18 giorni fa

692

modelcontextprotocol/python-sdk22k

perf: use deque for InMemoryTaskMessageQueue FIFO operations

Python19 giorni fa

75

mastra-ai/mastra22k

fix: prevent false [Circular] and [MaxDepth] in span serialization

TypeScript5 giorni fa

22153

PrefectHQ/prefect22k

perf: use deque for BFS queues in dbt freshness and settings scripts

Python19 giorni fa

64

PrefectHQ/prefect22k

fix: correct return type annotation for dbt_runner_freshness_success fixture

Python18 giorni fa

11

openai/openai-agents-python20k

perf: use deque for BFS queue and voice task ordering

Python19 giorni fa

86

marimo-team/marimo20k

fix: strip stray <p></p> HTML tags from imported Jupyter markdown cells

Python6 giorni fa

2191

marimo-team/marimo20k

perf: use deque for BFS queue in pytest_changed dependency traversal

Python19 giorni fa

32

aio-libs/aiohttp16k

perf: use deque for FormData.add_fields() queue

Python19 giorni fa

32

dottxt-ai/outlines14k

fix(llamacpp): correct EOS attention mask and vocab truncation bugs

Python18 giorni fa

1597

dottxt-ai/outlines14k

fix(dsl): add JSON double-quotes to string literals inside container types

Python18 giorni fa

3055

ag-ui-protocol/ag-ui13k

fix(python): align ReasoningMessageStartEvent.role with TypeScript SDK

Python19 giorni fa

22

ag-ui-protocol/ag-ui13k

fix(python): make Tool.parameters optional to match TypeScript SDK

Python19 giorni fa

11

modelcontextprotocol/typescript-sdk12k

fix: registerCapabilities registers logging/setLevel handler

TypeScript19 giorni fa

5210

Chainlit/chainlit12k

fix(data): prevent empty string from overwriting step content on upsert

Python18 giorni fa

232

livekit/agents9.7k

fix(anthropic): add dummy user message for Claude 4.6+ trailing assistant turns

Python18 giorni fa

222

livekit/agents9.7k

fix: prevent _cancel_speech_pause from poisoning subsequent user turns

Python6 giorni fa

123

github/copilot-sdk7.9k

fix(python): correct PermissionHandler.approve_all type annotations

TypeScript18 giorni fa

42

strands-agents/sdk-python5.3k

fix(openai): always use string content for tool messages

Python6 giorni fa

1517

strands-agents/sdk-python5.3k

fix(openai_responses): use output_text for assistant messages in multi-turn conversations

Python10 giorni fa

625

strands-agents/sdk-python5.3k

fix(graph): only evaluate outbound edges from completed nodes

Python12 giorni fa

442

griptape-ai/griptape2.5k

fix: use parameter value in output_schema validator error message

Python10 giorni fa

411

apache/hamilton2.4k

perf: use deque for BFS queue in graph traversal

Jupyter Notebook19 giorni fa

32

openlit/openlit2.3k

fix: make content truncation limits configurable via max_content_length

Python6 giorni fa

415226

weaviate/weaviate-python-client217

perf: use deque for iterator object cache

Python19 giorni fa

99

ben-vargas/ai-sdk-provider-gemini-cli69

feat: add thoughtSignature support for Gemini 3 models

TypeScript2 mesi fa

3315

Dashboard

Pull Requests

fix(openai): add type: message to Responses API input items

fix(langchain): normalize custom detector output to prevent KeyError in hash/mask strategies

fix: resolve chat template names before kwargs detection

fix(readingorder): assign FURNITURE content_layer to footer/header in container groups

fix: add missing 'json' keyword to graph memory prompts (fixes #4248)

fix(bedrock): group parallel tool results in single user message

fix(telemetry): skip signal handler registration in non-main threads

preserve file extension when filename stem is stripped entirely in gr.File

perf: use deque for SSE pending message queues in gradio_client

fix(events): correct type hints for gr.on triggers parameter

perf: use deque for FIFO queues in sequence parallel, superoffload, and compile

fix: exclude gpt-5.2-chat from temperature passthrough

fix(anthropic): handle OAuth tokens in count_tokens endpoint

fix(azure): forward realtime_protocol from config and relax api_version check for GA path

fix(tools): gracefully repair truncated JSON in tool call arguments

fix: WebSearch interception fails with thinking enabled + SpendLog dedup

fix(streaming): prevent Vertex AI Claude content truncation when finish_reason races content

fix(bedrock): strip output_config from Bedrock Invoke requests

fix(cost_calc): update custom_llm_provider when base_model has different provider prefix

feat(anthropic): support top-level cache_control for automatic prompt caching

fix(featherless_ai): use correct FEATHERLESS_AI_API_KEY env var name

fix(models): set gpt-5.4-pro mode to responses — fixes #23014

fix(anthropic): populate output_config when reasoning_effort is used on Claude 4.6

fix: wrap input as list for non-default Azure models

fix: prevent deadlock in parallel MCP tool calls with header_provider (fixes #6094)

perf: use deque for BFS queue in get_knowledge_subgraph()

fix: guard against non-object tool arguments from LLMs

fix(react): replace static KaTeX CSS import with dynamic runtime injection

Python: fix(python/google): preserve thought_signature in Gemini function call parts

Add authentication support to OTLP exporter headers

fix(#469): deduplicate cursor-interactive elements in snapshot -C

fix: correct misleading output for cookies clear and tab close

fix(oauth): include client_id in token request body for client_secret_post

perf: use deque for InMemoryTaskMessageQueue FIFO operations

fix: prevent false [Circular] and [MaxDepth] in span serialization

perf: use deque for BFS queues in dbt freshness and settings scripts

fix: correct return type annotation for dbt_runner_freshness_success fixture

perf: use deque for BFS queue and voice task ordering

fix: strip stray <p></p> HTML tags from imported Jupyter markdown cells

perf: use deque for BFS queue in pytest_changed dependency traversal

perf: use deque for FormData.add_fields() queue

fix(llamacpp): correct EOS attention mask and vocab truncation bugs

fix(dsl): add JSON double-quotes to string literals inside container types

fix(python): align ReasoningMessageStartEvent.role with TypeScript SDK

fix(python): make Tool.parameters optional to match TypeScript SDK

fix: registerCapabilities registers logging/setLevel handler

fix(data): prevent empty string from overwriting step content on upsert

fix(anthropic): add dummy user message for Claude 4.6+ trailing assistant turns

fix: prevent _cancel_speech_pause from poisoning subsequent user turns

fix(python): correct PermissionHandler.approve_all type annotations

fix(openai): always use string content for tool messages

fix(openai_responses): use output_text for assistant messages in multi-turn conversations

fix(graph): only evaluate outbound edges from completed nodes

fix: use parameter value in output_schema validator error message

perf: use deque for BFS queue in graph traversal

fix: make content truncation limits configurable via max_content_length

perf: use deque for iterator object cache

feat: add thoughtSignature support for Gemini 3 models

Dashboard

Pull Requests

fix(openai): add type: message to Responses API input items

fix(langchain): normalize custom detector output to prevent KeyError in hash/mask strategies

fix: resolve chat template names before kwargs detection

fix(readingorder): assign FURNITURE content_layer to footer/header in container groups

fix: add missing 'json' keyword to graph memory prompts (fixes #4248)

fix(bedrock): group parallel tool results in single user message

fix(telemetry): skip signal handler registration in non-main threads

preserve file extension when filename stem is stripped entirely in gr.File

perf: use deque for SSE pending message queues in gradio_client

fix(events): correct type hints for gr.on triggers parameter

perf: use deque for FIFO queues in sequence parallel, superoffload, and compile

fix: exclude gpt-5.2-chat from temperature passthrough

fix(anthropic): handle OAuth tokens in count_tokens endpoint

fix(azure): forward realtime_protocol from config and relax api_version check for GA path

fix(tools): gracefully repair truncated JSON in tool call arguments

fix: WebSearch interception fails with thinking enabled + SpendLog dedup

fix(streaming): prevent Vertex AI Claude content truncation when finish_reason races content

fix(bedrock): strip output_config from Bedrock Invoke requests