* feat(copilot): display tool calls immediately in loading state during streaming
Display tool calls in loading state as soon as they are parsed during
OpenAI streaming, rather than waiting until processToolCall is invoked.
Changes:
- parseOpenAICompletion: Track initialized tool calls and display them
immediately when we have complete tool info (id + function.name)
- processToolCall: Updated comment to clarify it merges with existing
loading state set during parsing
This provides better UX by showing tool execution progress progressively
as the stream is parsed.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* feat(copilot): display Anthropic tool calls immediately in loading state
Apply the same immediate tool call display pattern to Anthropic streaming
that was implemented for OpenAI.
Changes:
- parseAnthropicCompletion: Display tool calls immediately in loading state
when tool_use blocks are received in the message event
This ensures consistent UX across both OpenAI and Anthropic providers,
showing tool execution progress as soon as tool calls are detected.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* feat(copilot): show Anthropic tool calls even earlier with temp displays
Display temporary loading states for Anthropic tool calls as soon as
inputJson events are received (when tool input starts streaming), then
replace them with real tool displays when complete tool_use blocks
arrive in the message event.
Changes:
- ToolCallbacks: Added removeToolStatus method to clean up temp displays
- AIChatManager: Implemented removeToolStatus to remove tool messages
from displayMessages array
- anthropic.ts:
* Display temp tool on first inputJson event (earliest indicator)
* Flush pending text message before showing temp tool (proper ordering)
* Remove temp display when complete tool_use block arrives
* Replace with real tool display via preAction
This provides the earliest possible feedback for Anthropic tool calls,
showing loading states as soon as the model starts generating tool
inputs rather than waiting for complete blocks.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* cleaning
* cleaning
* cleaning
* fix icon
* nit
* handle error
* nit
---------
Co-authored-by: Claude <noreply@anthropic.com>
* feat(aichat): create reusable CustomAIPrompts component
Extract custom AI prompts UI into a reusable component that can be
used in both workspace settings and user settings. Component includes:
- AI mode selector with visual indicators for set prompts
- Textarea with character limit
- Customizable title, description, and hint messages
Co-authored-by: centdix <centdix@users.noreply.github.com>
* refactor(aichat): update workspace AISettings to use reusable component
Replace inline custom prompts UI with the reusable CustomAIPrompts
component. Add hint about user-level custom prompts being available
in account settings and how they combine with workspace prompts.
Co-authored-by: centdix <centdix@users.noreply.github.com>
* feat(aichat): add user-level custom AI prompts in account settings
Add collapsible section in user settings for custom AI prompts:
- Stored in localStorage (key: userCustomAIPrompts)
- Collapsible UI to save space
- Visual indicator when prompts are configured
- Hint about prompt combination with workspace settings
- Prompts apply across all workspaces for the user
Co-authored-by: centdix <centdix@users.noreply.github.com>
* feat(aichat): combine workspace and user custom prompts
Update AIChatManager to combine workspace-level and user-level custom
prompts. Prompts are combined in order: workspace first, then user.
Add helper functions in aiStore.ts:
- getUserCustomPrompts(): retrieves user prompts from localStorage
- getCombinedCustomPrompt(mode): combines workspace + user prompts
All AI modes (script, flow, navigator, ask, API) now use combined
prompts, allowing users to append their own instructions to workspace
settings across all workspaces.
Co-authored-by: centdix <centdix@users.noreply.github.com>
* fix: remove unused imports
Remove unused imports to fix svelte-check errors:
- Remove unused 'get' from svelte/store in AIChatManager
- Remove unused 'copilotInfo' from aiStore in AIChatManager
- Remove unused 'AIMode' from AISettings
Co-authored-by: centdix <centdix@users.noreply.github.com>
* simplify
* nit
* fix
* fix
---------
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
Co-authored-by: centdix <centdix@users.noreply.github.com>
* draft mcp client
* testing
* fix
* cleaning
* mcp resource in inputtransforms
* cleaning
* big cleaning
* cleaning
* no arc
* add utils file
* refactor tools
* add mcp actions
* draft frontend
* send arguments from backend
* better frontend
* cleaning
* use token for auth
* add logo
* rm
* fix
* fix
* chore: refactor mcp for ai agents (#6829)
* Add Tool enum for AIAgent with backward compatibility
- Created Tool enum that can be either Windmill (FlowModule) or Mcp (resource reference)
- Created McpToolRef struct to hold MCP resource path
- Implemented custom Deserialize for Tool with backward compatibility:
- New format: {type: 'windmill'|'mcp', ...}
- Old format: FlowModule objects (automatically wrapped in Tool::Windmill)
- Updated AIAgent to use Vec<Tool> instead of Vec<FlowModule>
- Updated FlowValue::traverse_leafs to handle Tool enum
- Backward compatible: old flows with Vec<FlowModule> will deserialize correctly
* Refactor AI executor to process Tool enum instead of extracting MCP from input_transforms
- Separate Windmill tools and MCP resource paths from tools list
- Process Windmill FlowModules into Tool definitions
- Load MCP tools from resource paths in Tool::Mcp variants
- Remove old logic that extracted mcp_resources from input_transforms
- Import FlowModule, remove unused InputTransform
- Fix type issues: use .as_str() for path and handle Option<bool> properly
* handle in args
* mcp as flowmodule
* frontend
* config for mcp
* simplify logic
* fix ai executor logic
* cleaning
* clean frontend
* fix
* better resource picker
* fix and styling
* add endpoint to fetch tools
* apply tool filtering
* fix name validation
* better ui
* use cache
* fix
* fix merge
* refactor: Separate MCP tools from FlowModule in AIAgent
- Add new AgentTool, ToolValue, and McpToolValue types
- Update AIAgent to use Vec<AgentTool> instead of Vec<FlowModule>
- Implement From traits for clean conversion between AgentTool and FlowModule
- Add backward compatibility via custom deserializer for AgentTool
- Simplify resolve_module logic by reusing existing resolve_modules function
- Update traverse_leafs to handle AgentTool structure
This refactoring separates MCP tools from FlowModule tools, making the
type system clearer and eliminating the need to treat MCP servers as
a special case of FlowModule.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* refactor: Update ai_executor and worker_lockfiles for AgentTool
- Update ai_executor.rs to handle new AgentTool structure
- Separate MCP tools from FlowModule tools using ToolValue enum
- Convert AgentTool to FlowModule for backward compatibility
- Add imports for AgentTool and ToolValue types
- Update worker_lockfiles.rs for lazy loading optimization
- Convert AgentTool <-> FlowModule in insert_flow_modules
- Preserve lazy loading for FlowModule tools via modules_node
- Keep MCP tools inline (lightweight, no need for lazy loading)
- Maintain backward compatibility with existing flows
This enables the lazy loading optimization for FlowModule tools while
keeping MCP tools inline, balancing performance and simplicity.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
* cleaning
* adapt frontend
* cleaning
* cleaning
* type fix
* cleaning
* fix back comp
* move mcp button position
* nit
* cleaning
* fix nested removal
* cleaning
* opti
* fix chat markdown display
* fix chat messages layout
* fix back comp frontend
* fix deserializer
* nit
* simpler serializer
* use if else
---------
Co-authored-by: Claude <noreply@anthropic.com>
* feat: Add back apply code button in CodeDisplay for non-diff-based providers
- Added apply button that shows only in script mode for non-diff-based providers
- Button allows applying code directly to the current editor
- Only shows for providers that don't support diff-based editing (excludes openai, anthropic, googleai, azure_openai)
Fixes#6799
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
* better
* not only for non diff providers
---------
Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
Co-authored-by: centdix <farhadg110@gmail.com>
* add toggle option + chat interface
* backend impl
* draft
* put info in schema
* Revert "backend impl"
This reverts commit c534eeb49986424e2c12e2c5642be4e17ba380d1.
* chat interface in flow input
* cleaning
* add logic for running flow + styling
* handle historic args
* fix frontend changes
* add tables
* add conv list
* add endpoints
* adapt frontend
* list message logic
* save message in db
* save response in db
* cleaning
* better migrations
* refresh on new conv
* better logic for messages
* nit
* genere conversation uuid from frontend
* store chat mode info in flow status
* better ui for chat
* collapse chat
* ui
* infinite scroll on convs
* infinite scroll on messages
* fix ui
* new chat entry on new
* cleaning
* change setting logic
* fix test logic from flow input
* move toggle to input
* add warning modal when enabling chat mode
* add summary and explanation on inline script
* add hint for chat mode on user_message desc
* show chat message instead of input in graph
* add warning for triggers
* one logo when not expanded
* use infinitelist for conversations
* add warning when deployment in progress
* full width button
* better icon for menu
* better input + nits
* put toggle in action
* use waitjob
* cleaning
* cleaning
* scroll on new + cleaning
* use enum
* fix logic
* full screen
* cleaning
* exit on updatesqlx error
* Update SQLx metadata
* fix
* cleaning
* add for wait result endpoint
* add missing drop
* delete cascade
* fix: use macro version of query_as in flow_conversations.rs
Use sqlx::query_as! macro instead of query_as function for compile-time
SQL validation and better type safety
Co-authored-by: centdix <centdix@users.noreply.github.com>
* fix: update comment to clarify conversation message update condition
The comment now accurately reflects that the update happens when
it's a flow and it's done (flow_is_done)
Co-authored-by: centdix <centdix@users.noreply.github.com>
* fix: only parse chat_input_enabled if conditions are met
Move the parse_chat_input_enabled() call inside the condition check
to avoid unnecessary parsing when the flow is not done or unsuccessful
Co-authored-by: centdix <centdix@users.noreply.github.com>
* fix: use the same transaction for conversation creation
Pass transaction to get_or_create_conversation_with_id instead of
creating a new one, ensuring all operations are atomic
Co-authored-by: centdix <centdix@users.noreply.github.com>
* fix: remove update trigger and handle updated_at in application code
Remove the database trigger that automatically updates conversation
timestamp and instead update it explicitly when creating messages.
This gives better control and consistency.
Co-authored-by: centdix <centdix@users.noreply.github.com>
* Update SQLx metadata
* cleaning
* feat(aiagent): handle memory (#6719)
* implement memory
* s3 logic for memory
* fix typo
* much cleaner
* cleaning
* cleaning
* only if chat
* display nit
* nit
* fix stack overflow
* cleaning
* use len arg from input
* cleaning
* change order
* delete memory when conv deleted
* cleaning
* nit
* show description in expr mode
* opti
* opti
* updatee ref
* store string as simple string
* use markdown
* do not wait for deletion
* add delete loading
* fix logic
* fix markdown
* Update ee-repo-ref.txt
* Update SQLx metadata
* fix in test interface
* nit
* nit
* fix layout
* use memory_id to store memory
* shorter description
* rls + grant
* fix text overflow
* extract output from res
* cleaning
* handle streaming
* cleaning
* fix tool error
* nit
* update ref
* fix
* Update SQLx metadata
* nit
---------
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>
Co-authored-by: centdix <centdix@users.noreply.github.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
* Revert "use diffs based edits"
This reverts commit 4ef6bce562.
* feat(aichat): use diff-based edits for OpenAI/Anthropic providers, whole code for others
- Check the current model provider at runtime
- Use diff-based approach (with diffs array) for OpenAI and Anthropic
- Use whole code replacement for all other providers
- Update tool definition to support both parameters
- Update system prompt with conditional instructions based on provider
Co-authored-by: centdix <centdix@users.noreply.github.com>
* fix
* cleaning
* cleaning
* cleaning
* cleaning
---------
Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>
Co-authored-by: centdix <centdix@users.noreply.github.com>
* feat: ai agent steps base
* better backend and graph
* feat: anthropic, log viewer
* nit
* fix(frontend): hide tool nodes from timeline
* move ai agent actions from flow status to flow status module
* nits and workspace/hub scripts support
* tmp ref
* fix merge
* feat: display agent tools status in the graph
* fix reactivity
* fix flow status
* nit
* Remove inline type restriction from TypeScript AI script generation
Remove the requirement to inline object types in TypeScript AI chat prompts. The AI will now only receive guidance about using RT.ResourceType for resource types, without being forced to inline other parameter types.
Fixes#6099
Co-authored-by: Ruben Fiszel <rubenfiszel@users.noreply.github.com>
* Rename TS_INLINE_TYPE_INSTRUCTION to TS_RESOURCE_TYPE_INSTRUCTION
Co-authored-by: centdix <centdix@users.noreply.github.com>
* remove
---------
Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>
Co-authored-by: Ruben Fiszel <rubenfiszel@users.noreply.github.com>
Co-authored-by: centdix <centdix@users.noreply.github.com>
Co-authored-by: centdix <farhadg110@gmail.com>
* feat(aichat): add mode-specific placeholders for better user guidance
- Replace generic "Ask anything" placeholder with mode-specific suggestions
- Script mode: "Modify this script, fix errors, or generate new code..."
- Flow mode: "Edit this flow, add steps, or modify workflow logic..."
- Navigator mode: "Help me navigate Windmill or find features..."
- API mode: "Make API calls to fetch data or manage resources..."
- Ask mode: "Ask questions about Windmill features and documentation..."
- Maintains backward compatibility for custom placeholders
- Uses Svelte 5 reactive $derived.by for dynamic updates
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
* feat(aichat): make placeholder prop optional and simplify logic
- Make placeholder prop optional instead of defaulting to 'Ask anything'
- Remove !== 'ask anything' logic check, now simply checks if placeholder is provided
- When placeholder is provided, use it directly; otherwise fall back to mode-specific placeholders
- Maintains backward compatibility while simplifying the conditional logic
Co-authored-by: centdix <centdix@users.noreply.github.com>
* feat(aichat): use AIMode enum instead of hardcoded strings
- Import AIMode enum in AIChatInput.svelte
- Replace hardcoded mode strings with enum values in switch statement
- Update focusInput function and template condition to use enum
- Maintains consistency with type-safe enum usage throughout codebase
Co-authored-by: centdix <centdix@users.noreply.github.com>
---------
Co-authored-by: claude[bot] <209825114+claude[bot]@users.noreply.github.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
Co-authored-by: centdix <centdix@users.noreply.github.com>
* add test script tool
* modify system prompt
* cleaning
* same for flows
* cleaning
* apply code when confirm test + fix circular dep
* cleaning
* factorize
* display error
* cleaning
* fix
* update comment
* prompts
* cleaner code
* show logs in separate container
* format
centdix
Guilhem
Diego Imbert
dieriba
Ruben Fiszel
hugocasa
