OmniAgent Control UI Roadmap¶

Feature roadmap for the OmniAgent web-based Control UI, inspired by OpenClaw's dashboard architecture.

Current Features (Implemented)¶

[x] Chat interface with message history
[x] Conversation sidebar with localStorage persistence
[x] Voice input (Web Speech API)
[x] Translation feature
[x] Chart rendering (ECharts + ChartIR)
[x] Dark theme support
[x] Stop button to cancel generation
[x] Phone number display for voice calls
[x] Model selection dropdown

v0.12.0 Additions¶

The following features were added in v0.12.0:

[x] Image Generation - OpenAI-compatible image endpoints via omniimage (DALL-E, FLUX)
[x] GitHub Skill - Search issues, PRs, and code across repositories
[x] Web Content Skill - Extract and process content from URLs
[x] Auto-Reply System - Automatic message responses with configurable triggers
[x] Setup Wizard - Interactive CLI setup with omniagent setup
[x] Doctor Command - Diagnose configuration and connectivity with omniagent doctor
[x] Sessions Command - Manage conversation sessions via CLI

Phase 1: Tools & Skills Panel¶

Status: Implemented (Core Features)

Display registered agent tools and their capabilities.

Features¶

[x] Tools sidebar panel (collapsible)
[x] List all registered tools with icons
[x] Tool descriptions and parameter schemas
[x] Tool detail modal with parameter info
[x] Tools grouped by category
[x] Tool usage statistics (call count, last used)
[ ] Tool invocation history in chat

API Endpoints¶

[x] GET /api/v1/tools - List available tools
[ ] GET /api/v1/tools/{name} - Get tool details

UI Components¶

[x] Tools panel toggle button in header (gear icon)
[x] Collapsible tools list (right sidebar)
[x] Tool detail modal/popover
[ ] Tool call indicators in chat messages

Implementation Files¶

api/openai/handler.go - Added ListTools to AgentHandler interface
api/openai/server.go - Added /api/v1/tools endpoint
openai/adapter.go - Implemented ListTools method
api/openai/web/index.html - Added tools panel UI

Phase 2: Reasoning & Tool Call Streams¶

Status: Implemented (Thinking Blocks)

Show the agent's thinking process and tool calls during responses.

Features¶

[x] Collapsible "Thinking" sections in assistant messages
[x] Real-time thinking block display during streaming
[x] Animated "Thinking..." indicator while processing
[x] Parse <thinking> and <think> tags
[x] Tool call visualization with expand/collapse
[x] Tool arguments display (formatted JSON)
[ ] Tool result preview
[ ] Execution timeline

Implementation¶

[x] Parse <thinking> tags from responses (both streaming and completed)
[x] Render thinking blocks with collapsible UI
[x] Purple accent color for thinking blocks
[x] Detect tool_calls in streaming responses
[x] Render tool calls as expandable cards
[ ] Show tool results inline or in modal

UI Features¶

Thinking blocks display with brain icon
Real-time streaming inside thinking blocks
"Thinking..." animates while processing, changes to "Thought" when complete
Click to expand/collapse thinking content
Proper handling of mixed thinking + regular content

Phase 3: System Status Bar¶

Status: Implemented

Real-time system status and health monitoring.

Features¶

[x] Connection status indicator (connected/disconnected)
[x] Current model info display
[x] Token usage counter (current session)
[x] Response time metrics
[x] Health check polling (every 30s)
[x] Gateway version display

API Endpoints¶

[x] GET /health - Health check (already exists)
[x] GET /api/v1/status - Detailed status including version

UI Components¶

[x] Status bar at bottom of input area
[x] Green/red connection indicator dot
[x] Model name display (synced with model select)
[x] Token count with icon
[x] Response time display

Implementation¶

Health check polling every 30 seconds
Token estimation (~4 chars/token)
Response time tracked per message
Token count updates on:
Sending messages
Receiving responses
Switching conversations
Creating new conversations

Phase 4: Settings Panel¶

Status: Implemented

User-configurable agent settings from the UI.

Features¶

[x] Settings modal/drawer
[x] Temperature slider (0.0 - 2.0)
[x] Max tokens input
[x] System prompt editor
[x] API key configuration (masked input)
[x] Theme toggle (light/dark/system)
[x] Voice settings (language selection)
[ ] Voice continuous mode toggle

Persistence¶

[x] Store settings in localStorage
[ ] Sync to server via API (optional)

UI Components¶

[x] Settings sliders icon in header
[x] Modal with grouped sections
[x] Form validation
[x] Reset to defaults button
[x] Apply settings without reload

Implementation¶

Settings stored in omniagent_settings localStorage key
Default values: temperature=0.7, maxTokens=4096, theme=system
Theme applies immediately (light/dark/system with media query)
API calls include temperature, maxTokens, and optional apiKey
System prompt prepended to messages when set

Phase 5: Activity Log¶

Status: Implemented

Real-time event stream showing agent activity.

Features¶

[x] Expandable activity log panel (bottom drawer)
[x] Event types: chat, tool, error, system
[x] Timestamp and duration display
[x] Filter by event type
[x] Search within logs
[x] Export logs (JSON format)
[x] Clear log functionality

Event Types¶

CHAT    - User messages, assistant responses
TOOL    - Tool loading events
ERROR   - Connection failures, request errors
SYSTEM  - App startup, settings changes, generation stops

UI Components¶

[x] Toggle button in header (clipboard icon)
[x] Bottom drawer panel (collapsible)
[x] Filter buttons (All/Chat/Tools/System/Errors)
[x] Search input with real-time filtering
[x] Export and clear buttons
[x] Color-coded event types with badges

Implementation¶

Activity events stored in memory (max 500 events)
Relative timestamps ("Just now", "5m ago", "2h ago")
Duration tracking for response times
Events logged for:
User message sent
Assistant response received (with tokens and duration)
Generation stopped
Connection status changes
Tools loaded
Settings saved
App initialization

Phase 6: Enhanced Session Management¶

Status: Implemented

Advanced conversation/session features.

Features¶

[x] Session metadata display (created, message count, tokens)
[x] Export conversation as markdown
[x] Export conversation as JSON
[x] Session archival (hide without delete)
[x] Session duplication (copy full conversation)
[x] Session forking (fork from message) - infrastructure ready
[x] Session search (across title, tags, and messages)
[x] Session tags/labels (add, remove, display)
[x] Session title editing

UI Components¶

[x] Session info modal (metadata, tags, export, actions)
[x] Search bar in sidebar
[x] Toggle archived conversations
[x] Tags display in conversation list
[x] Fork indicator for forked conversations
[x] Message count in conversation list

Implementation¶

Client-side localStorage persistence
Search filters title, tags, and message content
Archived sessions hidden by default (toggle to show)
Tags stored as array in conversation object
Export formats: Markdown (readable) and JSON (structured)
Duplicate creates full copy with "(Copy)" suffix
Fork ready for per-message branching UI

Phase 7: Cron/Scheduled Jobs¶

Status: Implemented

Schedule automated agent tasks.

Features¶

[x] List scheduled jobs
[x] Create/edit/delete jobs
[x] Cron expression builder (with presets)
[x] Run job manually (trigger)
[x] Enable/disable jobs
[x] Interval and one-time scheduling
[ ] Job execution history (future)

Action Types¶

send_message - Send a message to an agent session
call_webhook - Make HTTP request to a webhook URL
call_tool - Invoke a registered agent tool

API Endpoints¶

GET /api/v1/cron/jobs - List all scheduled jobs
POST /api/v1/cron/jobs - Create a new job
GET /api/v1/cron/jobs/{id} - Get job details
PUT /api/v1/cron/jobs/{id} - Update a job
DELETE /api/v1/cron/jobs/{id} - Delete a job
POST /api/v1/cron/jobs/{id}/trigger - Run job immediately
POST /api/v1/cron/jobs/{id}/enable - Enable a job
POST /api/v1/cron/jobs/{id}/disable - Disable a job

UI Components¶

[x] Cron toggle button in header (clock icon)
[x] Collapsible cron jobs panel (right sidebar)
[x] Job cards with status, schedule, actions
[x] Create/edit modal with:
Schedule type tabs (Cron/Interval/Once)
Cron expression presets
Action type selector
Dynamic action fields
[x] Quick actions (trigger, enable/disable, delete)

Implementation Files¶

cron/executor.go - Job action execution handler
cron/tools.go - Cron skill with AgentAware interface
api/openai/handler.go - Cron types and AgentHandler interface
api/openai/server.go - REST API endpoints
openai/adapter.go - AgentHandler implementation
api/openai/web/index.html - Web UI cron panel

Phase 8: Multi-Agent Support¶

Status: Implemented

Support for multiple agent configurations.

Features¶

[x] Agent switcher dropdown (via model selector)
[x] Agent configuration editor (create/edit modal)
[x] Per-agent tool assignments (allowed_tools, denied_tools)
[x] Agent-specific system prompts
[x] Agent cloning
[x] Agent registry with KVS persistence
[x] Multi-agent adapter for request routing
[x] REST API for agent management

API Endpoints¶

GET /api/v1/agents - List all configured agents
POST /api/v1/agents - Create a new agent
GET /api/v1/agents/{id} - Get agent details
PUT /api/v1/agents/{id} - Update an agent
DELETE /api/v1/agents/{id} - Delete an agent
POST /api/v1/agents/{id}/clone - Clone an agent

Configuration¶

Agents can be configured via YAML:

# Single agent (backward compatible)
agent:
  provider: anthropic
  model: claude-sonnet-4
  api_key: ${ANTHROPIC_API_KEY}

# Multiple agents
agents:
  - id: default
    name: General Assistant
    provider: anthropic
    model: claude-sonnet-4
    system_prompt: "You are a helpful assistant."

  - id: coder
    name: Code Expert
    provider: anthropic
    model: claude-sonnet-4
    system_prompt: "You are an expert programmer..."
    allowed_tools: [read, write, edit, glob, grep]

UI Components¶

[x] Agents toggle button in header (users icon)
[x] Collapsible agents panel (right sidebar)
[x] Agent cards with name, model, status
[x] Create/Edit modal with form fields
[x] Quick actions (switch, edit, clone, delete)

Implementation Files¶

agent/registry/config.go - AgentConfig type
agent/registry/store.go - KVS persistence
agent/registry/registry.go - Registry management
agent/registry/registry_test.go - Registry unit tests
config/config.go - Config enhancements
openai/multi_adapter.go - Multi-agent routing
openai/multi_adapter_test.go - Multi-agent adapter tests
openai/cron_handler.go - Shared cron handler (embedded by adapters)
api/openai/handler.go - Agent types and interface
api/openai/server.go - REST API endpoints
api/openai/server_test.go - API handler tests
cmd/omniagent/commands/openai.go - CLI integration
cmd/omniagent/commands/gateway.go - Gateway integration
api/openai/web/index.html - Web UI agents panel

Phase 9: Usage Analytics¶

Status: Implemented

Token usage tracking, cost estimation, and visualization.

Features¶

[x] Token usage tracking per request
[x] Cost estimation based on model pricing
[x] Latency metrics
[x] Time-series data aggregation (hourly/daily)
[x] Summary statistics (total requests, tokens, cost)
[x] Usage breakdown by model
[x] ECharts visualizations (tokens, cost, model distribution)
[x] Recent requests history table

API Endpoints¶

GET /api/v1/usage - Usage summary for time range
GET /api/v1/usage/timeseries - Time-bucketed usage data
GET /api/v1/usage/records - Recent usage records
GET /api/v1/usage/chart/tokens - Token chart (ChartIR format)
GET /api/v1/usage/chart/cost - Cost chart (ChartIR format)
GET /api/v1/usage/chart/models - Model distribution chart (ChartIR format)

Query Parameters¶

since - Start time (RFC3339 or relative: 1h, 24h, 7d, 30d)
until - End time (RFC3339, defaults to now)
interval - Bucket size (minute, hour, day)
limit - Max records for /records endpoint

Model Pricing¶

Default pricing for cost estimation:

Model	Input $/1M tokens	Output $/1M tokens
claude-sonnet-4*	3.00	15.00
claude-opus-4*	15.00	75.00
claude-haiku-4*	0.25	1.25
gpt-4o*	2.50	10.00
gpt-4o-mini*	0.15	0.60
default	1.00	2.00

UI Components¶

[x] Usage toggle button in header (chart icon)
[x] Usage analytics modal with:
Time range selector (1h, 24h, 7d, 30d)
Summary cards (requests, tokens, cost, latency)
Token usage stacked bar chart
Cost line chart with area fill
Model distribution pie chart
Model breakdown table
Recent requests table

Implementation Files¶

api/openai/usage.go - Usage types, store, and chart generation
api/openai/handler.go - UsageHandler interface
api/openai/server.go - Usage API endpoints
api/openai/streaming.go - Usage tracking in chat handler
api/openai/web/index.html - Usage analytics UI

Dependencies¶

github.com/grokify/echartify/chartir - ECharts IR generation

Phase 10: Memory/Context Panel¶

Status: Implemented

Semantic memory for storing and retrieving information across conversations.

Features¶

[x] Memory panel (right sidebar)
[x] Collection selector
[x] Semantic search with query input
[x] Memory list with content, key, metadata
[x] Add memory modal
[x] Delete memory functionality
[x] Collection refresh

API Endpoints¶

GET /api/v1/memories - List memories in collection
POST /api/v1/memories - Store new memory
GET /api/v1/memories/search - Semantic search
DELETE /api/v1/memories/{key} - Delete memory
GET /api/v1/memories/collections - List collections

Query Parameters¶

collection - Collection name (default: "default")
q - Search query (for /search endpoint)
limit - Max results

UI Components¶

[x] Memory toggle button in header (lightbulb icon)
[x] Collection dropdown with refresh button
[x] Search input with Enter to search
[x] Memory cards with content, key, metadata, score
[x] Add memory modal with content, key, collection, metadata
[x] Delete confirmation

Integration¶

The memory panel connects to the omnimemory library which provides:

Short-term memory (conversation context)
Long-term memory (persistent knowledge)
Vector embeddings for semantic search
Collection-based organization
Metadata support
KVS persistence

Implementation Files¶

api/openai/handler.go - MemoryHandler interface and types
api/openai/server.go - Memory API endpoints
api/openai/web/index.html - Memory panel UI
skills/compiled/memory/memory.go - Memory skill (omnimemory integration)

Technical Architecture¶

API Router Stack (Chi + Huma + ogen)¶

The API server uses a hybrid architecture combining three routing technologies:

Chi Router (base)
├── /openai/v1/chat/completions  → StreamingHandler → ogen (SSE streaming)
├── /openai/v1/models            → ogen handler
├── /openai/v1/models/{model}    → ogen handler
├── /openai/v1/images/*          → Huma (image generation via omniimage)
├── /api/v1/tools                → Huma (auto-generated OpenAPI)
├── /api/v1/agents/*             → Huma
├── /api/v1/cron/jobs/*          → Huma
├── /api/v1/usage/*              → Huma
├── /api/v1/memories/*           → Huma
├── /api/health                  → Huma
├── /api/openapi.json            → Merged spec (ogen input + Huma generated)
├── /docs                        → Scalar UI
└── /                            → Web UI (static)

Component Responsibilities¶

Chi Router (github.com/go-chi/chi/v5): Base HTTP router with middleware support (RealIP, Recoverer). Provides clean path-based routing and compatibility with standard http.Handler.
ogen (github.com/ogen-go/ogen): Handles OpenAI-compatible endpoints (/openai/v1/chat/completions, /openai/v1/models). Used for SSE streaming support and type-safe request/response handling generated from OpenAPI spec.
Huma (github.com/danielgtaylor/huma/v2): Handles OmniAgent extension endpoints (tools, agents, cron, usage, memory, health). Provides automatic OpenAPI 3.1 schema generation from Go types with validation.
Scalar (github.com/bdpiprava/scalar-go): Serves interactive API documentation at /docs with modern UI, dark mode, and search.

OpenAPI Spec Merging¶

The /openapi.json endpoint serves a merged OpenAPI specification that combines:

ogen input spec (openapi/openai-minimal.yaml): OpenAI-compatible chat completion and models endpoints
Huma-generated spec: Auto-generated from registered Huma operations (tools, agents, cron, usage, memory, health)

This provides a complete, accurate API reference for all endpoints.

Implementation Files¶

api/openai/server.go - Chi router setup, Huma integration, route mounting
api/openai/operations/*.go - Huma operation definitions (types, tools, agents, cron, usage, memory, health, images)
api/openai/image_handler.go - ImageHandler interface and omniimage integration
api/openai/openapi_merge.go - OpenAPI spec merging logic
api/openai/docs.go - Scalar documentation handler
api/openai/streaming.go - SSE streaming handler for chat completions
api/openai/internal/ogen/ - Generated ogen code from OpenAPI spec

Frontend Stack¶

Vanilla JavaScript (current)
Tailwind CSS for styling
ECharts for visualization
Web Speech API for voice

Backend Integration¶

OpenAI-compatible API endpoints
Server-Sent Events (SSE) for streaming
WebSocket for real-time updates (future)

State Management¶

localStorage for persistence
In-memory state for session
Server sync for settings (optional)

OmniAgent Control UI Roadmap¶

Current Features (Implemented)¶

v0.12.0 Additions¶

Phase 1: Tools & Skills Panel¶

Features¶

API Endpoints¶

UI Components¶

Implementation Files¶

Phase 2: Reasoning & Tool Call Streams¶

Features¶

Implementation¶

UI Features¶

Phase 3: System Status Bar¶

Features¶

API Endpoints¶

UI Components¶

Implementation¶

Phase 4: Settings Panel¶

Features¶

Persistence¶

UI Components¶

Implementation¶

Phase 5: Activity Log¶

Features¶

Event Types¶

UI Components¶

Implementation¶

Phase 6: Enhanced Session Management¶

Features¶

UI Components¶

Implementation¶

Phase 7: Cron/Scheduled Jobs¶

Features¶

Action Types¶

API Endpoints¶

UI Components¶

Implementation Files¶

Phase 8: Multi-Agent Support¶

Features¶

API Endpoints¶

Configuration¶

UI Components¶

Implementation Files¶

Phase 9: Usage Analytics¶

Features¶

API Endpoints¶

Query Parameters¶

Model Pricing¶

UI Components¶

Implementation Files¶

Dependencies¶

Phase 10: Memory/Context Panel¶

Features¶

API Endpoints¶

Query Parameters¶

UI Components¶

Integration¶

Implementation Files¶

Technical Architecture¶

API Router Stack (Chi + Huma + ogen)¶

Component Responsibilities¶

OpenAPI Spec Merging¶

Implementation Files¶

Frontend Stack¶

Backend Integration¶

State Management¶

References¶