v2.2.49

What's Changed

• fix: replace prefix-based model detection with capability-based detection by @MervinPraison in #803
• fix: resolve SyntaxError in autoagents.py by @MervinPraison in #817

Full Changelog: v2.2.48...v2.2.49

v2.2.48

What's Changed

• fix: handle JSON parsing for Gemini self-reflection by @MervinPraison in #800
• fix: update AutoAgents prompt to generate proper TaskConfig objects by @MervinPraison in #801

Full Changelog: v2.2.47...v2.2.48

v2.2.47

What's Changed

• fix: add multi-provider support for AutoAgents by @MervinPraison in #791
• fix: prevent double output when using litellm models by @MervinPraison in #789
• refactor: extract debug logging into centralized method by @MervinPraison in #793
• PR #791: Changes from Claude by @MervinPraison in #794
• refactor: eliminate sync/async duplication in llm.py by @MervinPraison in #781

Full Changelog: v2.2.46...v2.2.47

v2.2.46

What's Changed

• refactor: extract streaming chunk processor to eliminate code duplication by @MervinPraison in #782

Full Changelog: v2.2.45...v2.2.46

v2.2.45

What's Changed

• refactor: extract Ollama handler to eliminate code duplication by @MervinPraison in #783

Full Changelog: v2.2.44...v2.2.45

v2.2.44

What's Changed

• Add Air Quality Analysis Agent Notebook to Health Fitness Agents by @Dhivya-Bharathy in #775
• Add Financial Coaching Agent Notebook to Finance Market Job Agents Folder by @Dhivya-Bharathy in #776
• Add News and Podcast Aggregator Agent Notebook to Social Media Content Agents Folder by @Dhivya-Bharathy in #777
• Add Game Playing Agent (Tic Tac Toe) Notebook to Programming Code Analysis Agents Folder by @Dhivya-Bharathy in #778
• Add 5 domain-specific AI agent examples to exa-tool directory by @Dhivya-Bharathy in #780
• Add Domain Deep Research Agent Notebook to Research Knowledge QA Agents Folder by @Dhivya-Bharathy in #779
• fix: remove OPENAI_API_KEY requirement for non-OpenAI providers by @MervinPraison in #785
• Add grok4_intelligent_agents Folder to Exa Tool with 5 Agent scripts by @Dhivya-Bharathy in #786

Full Changelog: v2.2.43...v2.2.44

v2.2.43

What's Changed

• fix: add array schema validation to llm.py for MCP tools compatibility by @MervinPraison in #774

Full Changelog: v2.2.42...v2.2.43

v2.2.42

What's Changed

• refactor: make openai_client.py self-contained like llm.py by @MervinPraison in #771
• fix: add items property to array parameters in OpenAI tool definitions by @MervinPraison in #772
• fix: comprehensive array schema validation for MCP tools by @MervinPraison in #773

Full Changelog: v2.2.41...v2.2.42

v2.2.41

What's Changed

• refactor: Sync agent.py with llm.py patterns - Phase 1 by @MervinPraison in #765
• refactor: separate OpenAI logic from agent.py into llm/openai_client.py by @MervinPraison in #766
• fix: handle MCP tools returning lists in agent.py by @MervinPraison in #768
• fix: handle both ToolCall dataclass and OpenAI object in tool execution by @MervinPraison in #769
• fix: convert ToolCall dataclass to dict for JSON serialization by @MervinPraison in #770

Full Changelog: v2.2.40...v2.2.41

v2.2.40

What's Changed

• refactor: Add full tool format support to async method by @MervinPraison in #764

Full Changelog: v2.2.39...v2.2.40