System Online
Admin: ...
Overview
Fleet Operations: 100+ account overview, connection health, publish queue, daily output, insights and learning coverage.
TODAY'S OPERATIONS
Active Fleet
—
0 total configured profiles
Published Today
—
Live Threads posts delivered
Schedule Progress
—
Daily cadence slots completed
Pending & Queued
—
Queue slots: — in-flight
Fleet Scale & Reach
Operational breakdown across connected, queued, and learning accounts
Connected Profiles
—
Graph API authorized (—)
Fleet Avg Views
—
Average post impressions reach
Learning Samples
—
Advisory training baseline (—)
Insight Posts
—
Tracked by Insight Worker
Attention Needed:
—
Publish Queue:
—
Published Cycles:
—
Daily Cadence Status
Loading schedule status...
Autonomous time-slot dispatcher operates based on per-account cooldown gap and daily post limits.
AI Writer Engine
LLM copywriting and generation parameters
Language Model
—
Provider Endpoint
OpenAI-compatible
Fleet Operations
Manage account tokens, tone presets, content identities, and runtime execution policies.
Daily Schedule
Operations queue for autonomous post distribution, timing windows, and recovery retry.
Planned
0
Total today
Completed
0
Published to Threads
Pending
0
Awaiting window
Missed
0
Window expired
Errors
0
Pipeline failures
| # | Account | Schedule Time | Status | Topic | Views | Interactions | URL POST |
|---|
Normal Topic Sources
External discovery feeds for Normal Scheduler. Add a source once, then optionally scope it to selected profiles in this account.
Sources
0
Healthy
0
Issues
0
Add a source
Auto detect supports web feeds plus social profile URLs when the required API credential is configured.
Paste a URL only. The server detects Threads, Facebook, X, TikTok, YouTube, Reddit, RSS, sitemap or web automatically.
Scope
Current account
Social accounts:
YouTube supported
Reddit supported
X API
Facebook HTML-first
Threads HTML-first
TikTok API
Source library
Edit only when a URL, label or profile scope changes.
Publish Operations
Campaign generation, multi-account distribution, draft queues, and audit history.
1
Source Affiliate Links
Enter product URLs (one per line) to parse into campaign targets
0 valid URLs
2
Mode & Distribution
Select execution pipeline and profile allocation strategy
Publish Mode
Distribution Strategy
3
Select Target Accounts
Select ready accounts with active tokens and valid content focus
Selected: 0
No active campaign
Start a campaign from Campaign Composer. Live Writer progress and final results will appear here.
| Account | Type | Content Preview | Created | Status | Actions |
|---|
| Date | Account | Mode | Product / Content | Views | Interactions | Status | Threads URL |
|---|
| Detected | Account | Commenter | Comment | Generated Reply | Intent | Status | Threads | Actions |
|---|
Insights workspace
Performance & Intelligence
Post reach metrics, engagement leaderboards, trends analytics, and learned patterns.
Published Today
0
Posts across active fleet
Total Posts
0
Lifetime monitored posts
Fleet Avg Views
—
Average views per post
Avg Interactions
—
Likes & replies per post
Fleet Performance Ranking
Account leaderboard ranked by 7-day average views, interactions, and consistency
| Rank | # | Account | Avg Views | Interactions | Today | Trend | Samples |
|---|
| # | Account | Posts | Avg Views | High Views | Low Views | Avg Interactions | High Interactions | Low Interactions |
|---|
Learned Performance & Baseline Models
Advisory training feedback, winner loop patterns, and statistical performance baselines
Multi-Arm Bandit Experiments
Exploration trials across tone variations, angles, and content formats
| # | Account | Exploration | Best Arm | Trials | Reward | Policy |
|---|
System Settings
Configure model parameters, training policies, publishing, and automation secrets.
Copywriter LLM Endpoint
Primary language model credentials and provider configuration
Provider architecture interface.
Target LLM identifier.
API endpoint prefix.
Leave empty to keep current secret.
Optional per-request limit. Blank uses the project's Writer default; the model may stop earlier.
Review when changing models. GLM-5.3-Flash cannot disable thinking.
Supported values vary by model. GLM-5.3-Flash accepts low, high or max; “None” is different from not sending the field.
Generation timeout window.
Retries per queued Writer job after the first attempt.
Maximum scheduled drafts generated in parallel.
Test Model
Reply Writer LLM Endpoint
Independent model for short Threads replies. It does not inherit the content Writer model.
When disabled, this account will not generate or publish automatic replies. Pending replies stay paused until it is enabled again.
Kept separate so reply generation can use a cheaper/faster model.
Model used only by the future Reply Engine.
Leave empty to keep current secret.
Reply-specific token cap. Lower values reduce cost for short comments.
Independent timeout for reply generation.
Telegram Notification
Receive private notifications only for profiles and jobs owned by your account.
Create your own Telegram bot, then paste its token here.
Open your bot, send /start, then use the private chat ID for that conversation.
Draft is ready to review in Web Admin.
Published URL and profile information.
Writer or publishing errors that need attention.
Threads profile connection becomes unusable.
Threads token expired, revoked, or requires action.
Daily summary scoped to your own profiles.
Training & Self-Learning System
Autonomous learning analysis and post-performance optimization loop
Enable autonomous training and learning cycles.
Model utilized for pattern analysis.
OpenAI-compatible training endpoint base.
Leave empty to keep current secret.
Minimum sample size before analysis runs.
Minimum cooldown between training cycles.
Training analysis execution timeout.
Sample limit per account analysis run.
Concurrent profile batch size.
Fleet Topic Focus & Sources
Assigned primary topics, avoid rules, and categories across profiles
| Account | Assigned Focus Topics | Action |
|---|
Prompt
Writing guidance is editable. System Rules are protected and applied automatically.
Final Prompt Simulation
Preview of Writing Prompt + protected System Rules + selected profile context sent to Writer.
@
Threads API
Connect your Meta Threads app and generate long-lived access tokens from authorization codes.
App configuration
API host used for Threads OAuth and Graph requests.
Your Meta app identifier for Threads OAuth.
Leave empty to keep current secret.
Must exactly match the redirect URI configured in your Meta app.
Long-Lived Token Converter
Paste one authorization code per line. The server exchanges each code and returns a long-lived Threads access token without storing the generated token.
Auth code → Long-lived token
Up to 100 codes per batch.
Advanced Publishing Configuration
Protected publishing limits and direct publish endpoints.
Protected hard ceiling applied by the publishing system.
Use your USER_PASS as the Bearer token. Admin integrations may use AFFILIATE_DIRECT_PUBLISH_API_KEY.
Status URL template for the request_id returned by Direct Publish. USER_PASS can only read owned profiles.
Admin
Manage user access and protected system security settings.
Invite Users
Create USER_PASS invite codes and assign a stable alias namespace to each person.
Display name shown for the invited user.
New profiles will be named alias_user_1, alias_user_2, ...
Maximum profiles this user can create. Set 0 to block new profile creation.
| Username | alias_user | Status | Profiles | Profile Limit | Actions |
|---|
System Security
Admin authentication, proxy diagnostics, and encryption keys.
Leave empty to keep the current password. Changing it signs out existing sessions.
Controlled echo endpoint used to verify per-profile proxy routing.
Leave empty to keep current secret.