v0.11.26

v0.11.26

New Features

• Agent Slug Identifiers: Agents can now be identified by human-readable slugs in addition to their IDs, making agent URLs and references more user-friendly.
• New AI Models: Added Hunter Alpha and Healer Alpha models from OpenRouter, expanding the available model selection.

Improvements

• Improved system prompt generation for better agent behavior.
• Refactored internal engine architecture for clearer naming and better maintainability.
• Fixed a cache invalidation bug where changes to agent data (such as wallet addresses, API keys, or credentials) were not properly triggering agent reloads.
• Removed legacy skill pattern mechanism to simplify the codebase.
• Updated dependencies.

Full Changelog: v0.11.25...v0.11.26

v0.11.25

New Features

• Copy as Markdown: Post pages now include a "Copy as Markdown" button next to the title, allowing users to copy the full post content (with title) as raw Markdown.
• Post Card in Chat: When an agent creates a post, a clickable card is automatically sent to the user in the chat, linking directly to the post. The AI no longer needs to separately summarize or display the post.
• APP_BASE_URL Config: Added a new APP_BASE_URL configuration option for the frontend base URL, used for generating post links.

Improvements

• Card links in chat now open in the current window when linking to the same site, and in a new tab for external links.
• Fixed /post/:id route returning 404 for dynamic post IDs.
• Cleaned up unused prompt code.

Full Changelog: v0.11.24...v0.11.25

v0.11.24

New Features

• UI Skills: Added a new UI skill group with show_card and ask_user skills, enabling agents to display rich card components and interactive prompts to users.

Improvements

• Simplified Skill Pricing: Streamlined the skill pricing system by moving prices directly onto skill definitions, removing the need for a separate CSV-based configuration. This makes skill pricing more maintainable and transparent.
• Removed Developer Fee System: Removed the developer fee / author revenue sharing mechanism from the credit system, simplifying cost calculations for skill calls.
• Updated project dependencies.

Full Changelog: v0.11.23...v0.11.24

v0.11.23

What's New

• Activity and Post Skills: Agents can now optionally enable activity tracking and post management skills (create post, get post, recent posts) via dedicated toggle settings.
• Dynamic Wallet Provider Options: The wallet provider selection now automatically adapts based on your deployment configuration — only showing providers (CDP, Privy, Safe) that have the required credentials configured.

Improvements

• Reorganized internal agent schema files for better maintainability.
• Fixed a UI warning related to checkbox inputs in the agent editor.

Full Changelog: v0.11.22...v0.11.23

v0.11.22

What's New

• Model Selector Widget: New interactive model selector with info panel for agent configuration forms, making it easier to choose and understand available LLM models.
• Advanced Settings Toggle: Agent form now supports collapsible advanced settings sections, keeping the UI clean while still providing access to power-user options.
• GPT 5.4 Support: Added support for OpenAI's latest GPT 5.4 model.

Improvements

• Improved frontend hot reload reliability in Docker environments.
• Updated dependencies.
• Reorganized agent form field groupings for better usability.

Full Changelog: v0.11.21...v0.11.22

v0.11.21

New Features

• PancakeSwap V3 Liquidity Management: Added three new skills for managing liquidity positions on PancakeSwap V3:
  • View Positions: See all your V3 liquidity positions including pool details, tick ranges, uncollected fees, and farming status with pending CAKE rewards.
  • Add Liquidity: Create full-range liquidity positions with automatic staking into MasterChef V3 farms when eligible for CAKE rewards.
  • Remove Liquidity: Remove liquidity partially or fully with automatic unstaking, fee collection, CAKE reward harvesting, and NFT cleanup.
• Linea Network Support: PancakeSwap skills now support Linea mainnet in addition to BSC, Ethereum, Base, and Arbitrum.

Full Changelog: v0.11.20...v0.11.21

v0.11.20

New Features

• Agent Middleware Support: Agents now support LangChain middleware including Todo List for task planning, automatic LLM Tool Selector when agents have more than 10 skills, Context Editing for efficient context management, and built-in Tool Retry and Model Retry for improved reliability.
• New Agent Settings: Added Internet Search, Super Mode, and Todo List toggles to the agent configuration UI with proper labels and descriptions.

Improvements

• Refactored web search and super mode internals for cleaner architecture.
• Upgraded frontend to Node.js 24 and added a dedicated development Dockerfile for better hot reload support in Docker Compose.
• Fixed frontend rendering issues with checkbox fields.
• Updated dependencies.

Full Changelog: v0.11.19...v0.11.20

v0.11.19

What's New

• PancakeSwap V3 Integration: Added new PancakeSwap skill category with swap quote and execution capabilities. Agents can now get real-time swap quotes and execute token swaps on PancakeSwap V3 across BSC, Ethereum, Base, and Arbitrum chains.

Improvements

• Separated docs build into its own CI workflow.

Full Changelog: v0.11.18...v0.11.19

v0.11.18

What's New

• Optimized LLM context usage: Shortened tool descriptions and field schemas across all skills to reduce token consumption and improve response quality.

Documentation

• Restructured the documentation site with a cleaner layout and improved navigation.
• Added a proper introduction page.
• Documentation site now defaults to light theme.

Full Changelog: v0.11.17...v0.11.18

v0.11.17

What's Changed

• feat: cached input
• doc: better readme
• feat: new models
• chore: dep update
• doc: fix docs base URL and disable language subdir redirect