Skip to content

Enthusiast v1.7.0

Choose a tag to compare

View all tags
@Davsooonowy Davsooonowy released this 17 Jun 13:24
v1.7.0
a144017
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Enthusiast v1.7.0 is out! This release introduces agentic execution - a new way of defining agent interaction, designed to facilitate AI integration into systems, where the tasks run autonomously (human-out-of-the-loop approach) and are executable via API with deterministic outcomes.

Besides agentic executions, the release adds two new agents, a native Anthropic LLM provider and a LangChain package upgrade. It also includes a range of conversation, LLM provider, and infrastructure improvements.

Highlights

  • Agentic Execution

    • Introduced the interface for agentic execution definitions, with built-in execution result validation and self-recovery mechanism.
    • Exposed the API interface for managing and triggering agentic executions.
    • Implemented example agentic execution definitions for existing agents.
    • Integrated agentic executions into the existing UI.

  • New Agents

    • Added Catalog Web Import agent for importing product catalog data by extracting data from web pages.
    • Added Invoice Scanning agent for extracting structured data from invoice documents.

  • New LLM Provider

    • Added Anthropic LLM Provider - native provider that makes Claude models a first-class option.

  • LangChain Major Version Upgrade

    • Upgraded LangChain from 0.3 to 1.2.
    • Refactored agent conversation core logic to match the new version.

Other Changes

  • Integrations

    • Extended the Medusa connector with stock management.

  • Agent Conversation

    • Introduced a compactor mechanism that iteratively summarizes conversation history, preserving context when messages are trimmed to fit the LLM token limit.
    • Surfaced agent tool calls as live status updates in the conversation view.
    • Other user experience improvements.

  • LLM Providers

    • Stability fixes across Google GenAI, OpenAI, and Ollama providers.
    • Improved embedding support for model-specific vector size constraints.

  • Infrastructure

    • Consolidated shared agent tools into the new enthusiast-agent-tools plugin.
    • Removed the legacy enthusiast-file-parsers plugin.

Full Changelog: v1.6.1...v1.7.0

Assets 2
Loading