Skip to content

Releases: agno-agi/agno

v1.4.7

13 May 14:55
@dirkbrnd dirkbrnd
v1.4.7
fa8c4df
Compare
Choose a tag to compare
v1.4.7

Changelog

New Features:

  • Azure OpenAI Tools: Added image generation via Dall-E via Azure AI Foundry.
  • OpenTelemetry Instrumentation: We have contributed to the OpenInference project and added an auto-instrumentor for Agno agents. This adds tracing instrumentation for Agno Agents for any OpenTelemetry-compatible observability provider. These include Arize, Langfuse and Langsmith. Examples added to illustrate how to use each one (here).
  • Evals Updates: Added logic to run accuracy evaluations with pre-generated answers and minor improvements for all evals classes.
  • Hybrid Search and Reranker for Milvus Vector DB: Added support for hybrid_search on Milvus.
  • MCP with Streamable-HTTP: Now supporting the streamable-HTTP transport for MCP servers.

Improvements:

  • Knowledge Filters Cookbook: Instead of storing the sample data locally, we now pull it from s3 at runtime to keep the forking of the repo as light as possible.

Bug Fixes:

  • Team Model State: Fixed issues related to state being shared between models on teams.
  • Concurrent Agent Runs: Fixed certain race-conditions related to running agents concurrently.

Breaking changes:

  • Evals Refactoring:
    • Our performance evaluation class has been renamed from PerfEval to PerformanceEval
    • Our accuracy evaluation class has new required fields: agent, prompt and expected_answer
  • Concurrent Agent Runs: We removed duplicate information from some events during streaming (stream=True). Individual events will have more relevant data now.

What's Changed

New Contributors

Full Changelog: v1.4.6...v1.5.0

Contributors

manuhortet, kkpalczewski, and 8 other contributors
Loading

v1.5.0

13 May 20:03
@dirkbrnd dirkbrnd
v.1.5.0
f810e2c
Compare
Choose a tag to compare
v1.5.0

Changelog

New Features:

  • Azure OpenAI Tools: Added image generation via Dall-E via Azure AI Foundry.
  • OpenTelemetry Instrumentation: We have contributed to the OpenInference project and added an auto-instrumentor for Agno agents. This adds tracing instrumentation for Agno Agents for any OpenTelemetry-compatible observability provider. These include Arize, Langfuse and Langsmith. Examples added to illustrate how to use each one (here).
  • Evals Updates: Added logic to run accuracy evaluations with pre-generated answers and minor improvements for all evals classes.
  • Hybrid Search and Reranker for Milvus Vector DB: Added support for hybrid_search on Milvus.
  • MCP with Streamable-HTTP: Now supporting the streamable-HTTP transport for MCP servers.

Improvements:

  • Knowledge Filters Cookbook: Instead of storing the sample data locally, we now pull it from s3 at runtime to keep the forking of the repo as light as possible.

Bug Fixes:

  • Team Model State: Fixed issues related to state being shared between models on teams.
  • Concurrent Agent Runs: Fixed certain race-conditions related to running agents concurrently.

Breaking changes:

  • Evals Refactoring:
    • Our performance evaluation class has been renamed from PerfEval to PerformanceEval
    • Our accuracy evaluation class has new required fields: agent, prompt and expected_answer
  • Concurrent Agent Runs: We removed duplicate information from some events during streaming (stream=True). Individual events will have more relevant data now.

What's Changed

New Contributors

Full Changelog: v1.4.6...v1.5.0

Contributors

manuhortet, kkpalczewski, and 8 other contributors
Loading

v1.4.6

10 May 21:38
@dirkbrnd dirkbrnd
v1.4.6
bf1462a
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.
Compare
Choose a tag to compare
v1.4.6

Changelog

New Features:

  • Cerebras Model Provider: Added Cerebras as a model provider.
  • Claude Web Search: Added support for Claude’s new web search tool.
  • Knowledge Base Metadata Filtering (Beta): Added support for filtering documents by metadata

    • Two Ways to Apply Filters:

      • Explicit Filtering: Pass filters directly to Agent or during run/query

        # Option 1: Filters on Agent initialization
agent = Agent(
					knowledge=knowledge_base, 
					knowledge_filters={"filter_1": "abc"}
				)
     
# Option 2: Filters on run execution
agent.run("Tell me about...", knowledge_filters={"filter_1": "abc"})

        See docs here

      • Agentic Filtering: Agent automatically detects and applies filters from user queries

        # Enable automatic filter detection
agent = Agent(
					knowledge=knowledge_base, 
					enable_agentic_knowledge_filters=True
				)
     
# Agent extracts filters from query
agent.run("Tell me about John Doe's experience...")

        See docs here

    • Two approaches for adding metadata to documents:

      1. During Knowledge Base Initialization:

        knowledge_base = PDFKnowledgeBase(path=[
     {
		     "path": "file_1.pdf", 
		     "metadata": {
				     "user_id": "abc"
				  }
		 },
		 {
		     "path": "file_2.pdf", 
		     "metadata": {
				     "user_id": "xyz"
				  }
		 }
])

      2. During Individual Document Loading:

        knowledge_base.load_document(
     path="file.pdf",
     metadata={"user_id": "abc"}
)

    • Compatibility

      • Knowledge Base Types: PDF, Text, DOCX, JSON, and PDF_URL
      • Vector Databases: Qdrant, LanceDB, and MongoDB

Improvements:

  • User and Session ID in Tools: Added current_user_id and current_session_id as default variables in session_data.

Bug Fixes:

  • Knowledge Base ID Clashes: Knowledge files with overlapping names (e.g., abc.-.xyz.pdf and abc.-.def.pdf) were being incorrectly identified due to the readers using formatted names as unique id which were getting uniqueness conflict. Introduced a unique ID for each document in all the readers using uuidv4() to ensure strict identification and prevent conflicts.

What's Changed

Full Changelog: v1.4.5...v1.4.6

Contributors

dirkbrnd, Mustafa-Esoofally, and 3 other contributors
Loading

v1.4.5

06 May 15:30
@dirkbrnd dirkbrnd
v1.4.5
8d44e42
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.
Compare
Choose a tag to compare
v1.4.5

Changelog

New Features:

  • Embedder Support via AWS Bedrock: AwsBedrockEmbedder has been added with a default embedding model of cohere.embed-multilingual-v3.
  • Gemini Video Generation Tool: Added video generation capabilities to GeminiTools.

Improvements:

  • Apify Revamp: Complete revamp of ApifyTools to make it completely compatible with Apify actors.

Bug Fixes:

  • Tools with Optional Parameters on Llama API: Fixed edge cases with functions

What's Changed

New Contributors

Full Changelog: v1.4.4...v1.4.5

Contributors

alpha-xone, dirkbrnd, and 3 other contributors
Loading

v1.4.4

04 May 07:21
@dirkbrnd dirkbrnd
v1.4.4
7af20cd
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.
Compare
Choose a tag to compare
v1.4.4

Changelog

New Features:

  • OpenAI File Support: Added support for File attached to prompts for agents with OpenAIChat models.

Improvements:

  • Llama API: Various improvements for Llama and LlamaOpenAI model classes including structured output and image input support
  • Async Custom Retriever: The retriever parameter can now be an async function to be used with agent.arun and agent.aprint_response.
  • Gemini Video URL Input: Added support for Video(url=...) for Gemini.

Bug Fixes:

  • OpenAI Responses o3 / o4 Tools: Fixed broken tool use for advanced reasoning models on OpenAIResponses.
  • MCP on CLI Support: Fixed support for MCPTools usage while calling agent.acli_app.

What's Changed

New Contributors

Full Changelog: v1.4.3...v1.4.4

Contributors

tiberiu92, leandrodamascena, and 10 other contributors
Loading

v1.4.3

30 Apr 13:23
@ysolanky ysolanky
v1.4.3
2e9a39e
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.
Compare
Choose a tag to compare
v1.4.3

Changelog

New Features:

  • Llama API: Added native SDK and OpenAI-like model classes.

Improvements:

  • Claude: Added support for AWS Session token for Claude
  • DynamoDB: Added support for AWS profile-based authentication.

Bug Fixes:

  • Session Metrics: Fix for session metrics showing up as 0
  • HF Embedder fix: Fixed Hugging Face Embedder

What's Changed

New Contributors

Full Changelog: v1.4.2...v1.4.3

Contributors

emerleite, lironesamoun, and 5 other contributors
Loading

v1.4.2

24 Apr 21:26
@dirkbrnd dirkbrnd
v1.4.2
a545ab7
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.
Compare
Choose a tag to compare
v1.4.2

Changelog

New Features:

  • MCP SSE Support: Added support for connecting to SSE MCP Servers.
  • Tool Hooks: You can now have a hook that is wrapped around all tool calls. This works for Toolkits and custom tools. See this example.
  • Team Session State: You can now manage a single state dictionary across a team leader and team members inside tools given to the team leader/members. See this example.
  • Cartesia Tool: Added support for Cartesia for text-to-speech capabilities.
  • Gemini Image Tools: Added a tool that uses Gemini models to generate images.
  • Groq Audio Tools: Added a tool that uses Groq models to translate, transcribe and generate audio.

Improvements:

  • PubmedTools Expanded Results: Added expanded result sets for PubmedTools .
  • Variety in Tool Results: Custom tools can now have any return type and it would be handled before being provided to the model.

Bug Fixes:

  • Teams Shared Model Bug: Fixed issues where a single model is used across team members. This should reduce tool call failures in team execution.

What's Changed

New Contributors

Full Changelog: v1.4.1...v1.4.2

Contributors

dsaad68, manthanguptaa, and 4 other contributors
Loading

v1.4.1

23 Apr 17:28
@ashpreetbedi ashpreetbedi
v1.4.1
1b0363d
Compare
Choose a tag to compare
v1.4.1

What's Changed

New Contributors

Full Changelog: v1.4.0...v1.4.1

Contributors

mkschreder, gusmmm, and 4 other contributors
Loading

v1.4.0

23 Apr 11:24
@dirkbrnd dirkbrnd
v1.4.0
21e7431
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.
Compare
Choose a tag to compare
v1.4.0

Changelog

New Features:

  • Memory Generally Available: We have made improvements and adjustments to how Agentic user memory management works. This is now out of beta and generally available. See these examples and these docs for more info.
  • OpenAI Tools: Added OpenAITools to enable text-to-speech and image generation through OpenAI’s APIs.
  • Zep Tools: Added ZepTools and AsyncZepTools to manage memories for your Agent using zep-cloud

Improvements:

  • Azure AI Foundry Reasoning: Added support for reasoning models via Azure AI Foundry. E.g. Deepseek-R1.
  • Include/Exclude Tools: Added include_tools and exclude_tools for all toolkits. This allows for selective enabling / disabling of tools inside toolkits, which is especially useful for larger toolkits.

Bug Fixes:

  • Gemini with Memory: Fixed issue with deepcopy when Gemini is used with Memory.

Breaking Changes:

  • Memory: Agents will now by default use an improved Memory instead of the now deprecated AgentMemory.
    • agent.memory.messages → run.messages for run in agent.memory.runs (or agent.get_messages_for_session())
    • create_user_memories → enable_user_memories and is now set on the Agent/Team directly.
    • create_session_summary → enable_session_summaries and is now set on the Agent/Team directly.

What's Changed

New Contributors

Full Changelog: v1.3.5...v1.4.0

Contributors

manuhortet, ashpreetbedi, and 6 other contributors
Loading

v1.3.5

21 Apr 17:55
@dirkbrnd dirkbrnd
v1.3.5
9818ca5
Compare
Choose a tag to compare
v1.3.5

Changelog

Improvements:

  • Further Async Vector DB Support: Support added for:
  • Reasoning on Agno Platform:
    • Added extensive support for reasoning on the Agno Platform. Go see your favourite reasoning agents in action!
    • Changes from SDK
      • Now sending events for the different types of reasoning and populate the reasoning_content on RunResponse for stream/non-stream, async/non-async
      • Unified json structure for all types of reasoning in Reasoning events
  • Google Caching Support: Added support for caching files and sending the cached content to Gemini.

Bug Fixes

  • Firecrawl Scrape: Fixed issues with non-serializable types for during Firecrawl execution. #2883

What's Changed

New Contributors

Full Changelog: v1.3.4...v1.3.5

Contributors

ashpreetbedi, alpha-xone, and 2 other contributors
Loading