c.Sahand—CorText
YOUR PRIVATE INTELLIGENCE SPACE

Good to have
you back.

Sign in to continue to your workspace.

Private connection · Single-user workspace
c.Sahand—CorText
NAVIGATION
S
Workspace ownerPersonal account
Chat
AI HUB
✳
Connect your AI providersAdd your Gemini, OpenRouter, AIPrime, CodeCraft, or custom OpenAI/Anthropic API keys to get started.
WEB CONVERSATIONS

Your Chats

c.
THINK WITHOUT LIMITS

A space for
what’s next.

Gemini, OpenRouter, AIPrime, CodeCraft, custom chatbots & multi-model orchestration.

✳ Google Gemini⇄ OpenRouter🤖 Pluggable Chatbots⌘ Multi-Model Orchestration
📌

New conversation

⌄
✳⌄
THE INTELLIGENCE LAYER

Models & providers

Use Uncheck All to bench models in bulk, then check only the models you want active in Chat, Orchestration, and Failovers.

✳
SELECTED PROVIDER

Provider

Click any model card below to test live generation. Uncheck "Enabled" on any model you want to hide or exclude from failovers.

Needs setup
Select a provider or add its API key in Settings.
API keys are stored securely on this server
MULTI-MODEL COLLABORATION ENGINE

AI orchestration

Coordinate multiple AI models across providers to plan, execute, review, and refine complex tasks in the background.

TEAM & TASK CONFIGURATION

Configure Collaboration Team

Assign your Coordinator, Worker(s), and optional QA Auditor. You can also assign a trained Chatbot Persona to any role.

PARTICIPANT ROSTER (EXECUTION ORDER)
RESILIENCE & AUTOMATIC FAILOVER OPTIONS
Runs persistently in background workers — safe to close browser or switch tabs.
QUEUED STAGE: QUEUED Round 0 / 5 Calls 0 / 15

Task title

Job ID: —
Coordinator is working… Waiting for model response
AI CONTRIBUTION & RELIABILITY BREAKDOWN 100%
Messages Exchanged 0 0 Coordinator · 0 Worker
Review Revisions 0 Rounds completed: 0
Failovers & Retries 0 Automatic model switches
Total Model Time 0.0 s Combined API compute
APPROVED DELIVERABLE

Final Collaboration Result 1 FILE

🔍 Independent Second-Opinion Audit Have another AI model inspect all generated project files for runtime bugs, missing timezones, or broken logic.
AUDIT VERDICT Auditor Model
View full auditor output & corrected code
GENERATED PROJECT FILES (CLICK TO VIEW OR DOWNLOAD):
SANDBOXED LIVE HTML PREVIEW
HUMAN-IN-THE-LOOP ITERATION

Request Changes or Add Features to This Project

Tell the AI team what to fix, change, or add. Use the checkboxes below to control what context is sent to the models.

Adds +3 collaboration rounds and preserves your existing project files.
Collaboration Tasks 0 TASKS
REUSABLE MEMORY

Memories

Durable facts remembered from earlier conversations. They are injected into answers automatically when relevant.

ANDROID APP

Mobile App

Connect the CorText Android app to this hub. The app talks to the hub, and the hub talks to the AIs — the models never see the app or its IP.

NOT ACTIVATED
Activate mobile app
Turn on the mobile API so the Android app can connect to this hub.
Direct mode — connect the app straight to this hub
In the app, choose Direct mode and enter this address. The token below is the only credential the app needs — copy it into the app's token field.
Regenerating invalidates the old token at once — update the app (and the relay) with the new one.
Middle server — connect through a second server
Activate the middle server, type its IP and choose a port. The hub builds a one-liner installer from these — paste it into the middle server's SSH and it connects back to this hub.
Connection status
Middle server Not connected
App Never connected
MOVE OR RECOVER

Backup

Take everything with you — chats, personas, skills, memories, settings — so you can reset this server or move to a new one and restore it all in one step.

Export — download a backup
One file with everything: all chats, personas, skills, memories and settings, plus the login and encryption secrets a fresh hub needs to restore them. Download it before you reset or move the server.
⚠️ This file is extremely sensitive — it contains your login password hash, the API-key encryption key and the mobile token. Keep it somewhere safe and never share it.
Import — restore from a backup
Restoring replaces everything on this hub — all current chats, personas, skills, memories and settings — with the contents of the backup file.
EXTERNAL CHATBOT & TELEGRAM LOGS

Bot Sessions

Conversations happening through your Telegram bots and external chatbot plugins.

LEARNED KNOWLEDGE BASE

Skills

Skills the AI learned from links via /learn. They are injected into answers automatically when relevant.

PLUGGABLE CHATBOTS

Personas

Specialised AI assistants with their own instructions. Pick one in Chat or in the Android app to talk to it. Advanced options (voice, models, Telegram) live in Settings → Chatbots & personas.

New Persona

Name it and write what it should do.

OBSERVABILITY

Usage & analytics

Understand model activity, response times and token usage.

IN DEVELOPMENT
Tokens used—Waiting for telemetry module
Requests—Waiting for telemetry module
Average latency—Response timing will appear here
ACTIVITY VISUALIZATION · COMING SOON
YOUR HUB, YOUR RULES

Settings

Configure providers, build pluggable chatbots & voice clones, and personalize your workspace.

Add Custom API Provider

Connect any endpoint supporting the OpenAI / ChatGPT or Anthropic API specification. Once saved, it becomes available across Chat, Models, and AI Orchestration.

Provider Connection

OpenAI Spec

Select a provider tab above to configure credentials and test models.

Not connected
Select a provider tab above, configure its connection, and click "Save Key & Load Models".

Available Models

Click any model card below to test whether it works. Toggle "Enabled" to include or exclude it from Chat & Orchestration.

Need an API key for this provider? Open Provider Console ↗

Pluggable Chatbots & Voice Clones

Create specialized AI assistants or train a chatbot to talk and act like you from your own Telegram/WhatsApp chat exports, with per-bot model allowlists.

Create Chatbot / Persona

Configure personality, reference knowledge, allowed models, and Telegram bot bindings.

🎙️ Voice & Style Trainer (Learn How You Speak) Upload a Telegram Desktop result.json, WhatsApp .txt chat export, or paste a conversation sample below. The trainer isolates your replies and distills your exact tone and phrasing.
Check the exact models this bot is allowed to use:

Workspace preferences

Personalize how your workspace looks and behaves.

Color themeSwitch between Dark Mode and crisp GitHub Light.
Main Chat allowed models Choose whether Main Chat can pick from all globally enabled models or only a curated subset.
Select which globally enabled models appear in Main Chat:
Persian & BiDi typographyVazirmatn font + automatic RTL/LTR mixed sentence alignment.
Enabled
Pluggable provider & chatbot registryGemini, OpenRouter, AIPrime, CodeCraft, custom APIs, and trainable personas.
Active

Security & access

Controls for your private, single-user installation.

Single-user authenticationLogin protected with a bcrypt password hash and persistent SQLite session.
Enabled
Server-side provider & bot keysAPI credentials and Telegram bot tokens are never stored in browser storage.
Enabled
Change login credentialsUse the SSH control center: sudo cortext
SSH menu
SAHAND-CORTEXT· PRIVATE MULTI-PROVIDER AI WORKSPACE v0.9.0