Version: 1.0.0 (v0.7.0) Last Updated: January 31, 2026

A comprehensive guide to all AI-powered features in nself-chat, including message summarization, smart search, intelligent bots, and auto-moderation.

1. Overview
2. Getting Started
3. Message Summarization
4. Smart Search
5. Bot Interactions
6. Auto-Moderation
7. Privacy & Data Usage
8. Opt-In/Opt-Out
9. Troubleshooting
10. FAQ

nself-chat includes four major AI-powered capabilities:

All AI features in nself-chat are designed with graceful degradation:

• With AI APIs configured: Full semantic understanding, advanced summaries, and intelligent detection
• Without AI APIs: Local fallbacks provide basic functionality without requiring external services
• User control: All AI features respect user preferences and privacy settings

To see which AI features are available in your workspace:

1. Click your profile in the top-right corner
2. Go to Settings > AI Features
3. View the status dashboard:

✅ Message Summarization: Enabled (OpenAI)
✅ Smart Search: Enabled (Semantic)
✅ Auto-Moderation: Enabled (AI-powered)
⚠️ Bot Responses: Basic (No AI configured)

If you're an admin setting up AI features:

1. Navigate to Admin Panel > AI Configuration
2. Choose your AI provider (OpenAI or Anthropic recommended)
3. Enter your API key (see Admin Guide)
4. Test the connection
5. Enable features for your workspace

Note: Regular users don't need to configure anything - AI features work automatically once enabled by admins.

Message summarization uses AI to condense long conversations into concise, readable summaries. This helps you:

• Catch up quickly after being away
• Review lengthy discussions without reading every message
• Understand thread context at a glance
• Extract key decisions from conversations

What it does: Provides a comprehensive overview of channel activity over a time period.

When to use:

• Starting your workday
• Reviewing what happened while you were offline
• Understanding channel activity patterns

How to access:

1. Open any channel
2. Click the ⚡ Summarize button in the top toolbar
3. Select "Channel Digest"
4. Choose time period (Last hour, Today, Last 7 days, Custom)

Example Output:

📊 Channel Digest: #engineering (Last 24 hours)

Summary:
The engineering team discussed the authentication bug affecting login
flows, decided to implement 2FA for all users, and reviewed the Q1
roadmap. Sarah reported fixing the database migration issue.

Key Points:
• Authentication bug fix deployed to staging
• 2FA implementation approved for Q1
• Database migration completed successfully
• Code review process updated

Topics: Authentication, Security, Database, Process Improvements
Participants: 8 members | Messages: 156

What it does: Summarizes a conversation thread to help you understand the discussion.

When to use:

• Before joining a long thread
• Reviewing decisions made in a thread
• Understanding context without reading all replies

How to access:

1. Hover over any message with replies
2. Click "View Thread"
3. Click the 📝 Summary button at the top of the thread panel

Example Output:

🧵 Thread Summary

The team discussed implementing dark mode. Three design approaches
were proposed. Final decision: Use CSS variables for maximum
flexibility. Alex will create a design system document by Friday.

Key Decisions:
• CSS variables approach approved
• Design system doc due Friday
• Color contrast must meet WCAG AA standards

Participants: 4 | Messages: 23

What it does: Shows you what you missed since you were last online.

When to use:

• Returning after time away
• Checking in on channels you don't actively monitor
• Quick status updates

How to access:

1. Automatically appears when you open a channel with new messages
2. Or click your profile > "Catch Up" to see all channels

Example Output:

👋 Welcome back! You missed 47 messages in 3 channels.

#engineering (23 messages):
• Bug fix for authentication deployed
• Q1 roadmap reviewed and approved

#design (15 messages):
• New brand guidelines published
• User research findings shared

#general (9 messages):
• Team lunch scheduled for Friday
• Office will be closed on Monday

What it does: Generates a brief 1-2 sentence summary of any selection of messages.

When to use:

• Quick context on recent messages
• TL;DR for lengthy discussions
• Sharing context with others

How to access:

1. Select multiple messages (Shift+Click)
2. Right-click and choose "Summarize Selection"

Example Output:

Brief Summary: Discussion about API rate limiting. Team decided to
implement exponential backoff and increase default limits for premium users.

Step-by-step:

1. Navigate to any channel or thread
2. Look for the summarization controls:
   • ⚡ Button in the toolbar (Channel Digest)
   • 📝 Button in thread header (Thread Summary)
   • Right-click menu on selected messages (Quick Summary)
3. Choose summary type and time range
4. Click "Generate Summary"
5. Wait 2-5 seconds for AI processing
6. Review and optionally "Copy" or "Share" the summary

Keyboard Shortcuts:

• Cmd/Ctrl + Shift + S - Summarize current view
• Cmd/Ctrl + Shift + D - Generate channel digest
• Esc - Close summary panel

Step-by-step:

1. Tap the ⋮ Menu icon in the top-right
2. Select "Summarize"
3. Choose summary type
4. Tap "Generate"
5. View summary and tap "Done" when finished

Tips for Mobile:

• Summaries are optimized for mobile screens
• Swipe down to dismiss
• Long-press summary to copy

Options available:

To customize:

1. Click Settings in the summary panel
2. Adjust options
3. Click "Regenerate" to apply changes

If you see "Summarization not available":

✅ Solution: The workspace administrator needs to configure AI settings. This is normal and the feature will work once configured.

If you see "Using basic summary":

ℹ️ What this means: AI is not configured, but you'll still get a basic statistical summary showing message counts, participants, and time ranges.

Smart Search uses AI to understand the meaning of your queries, not just match keywords. This means:

• Natural language: Search like you talk - "show me discussions about the login bug"
• Conceptual matching: Finds related content even with different wording
• Context awareness: Understands synonyms and paraphrasing
• Relevance ranking: Best results first, combining semantic similarity and recency

Query: authentication issues Finds: Messages about login problems, password resets, 2FA errors, session timeouts

Query: who can help with React? Finds: Messages from team members discussing React, frontend, components, hooks

Query: decisions about the API Finds: Messages containing conclusions, approvals, or choices regarding API design

Natural Language:

"What did Sarah say about the database migration?"
→ Finds Sarah's messages related to database, migration, schema changes

Time-Based:

"bug reports from last week"
→ Filters to messages from the past 7 days containing bug descriptions

Multi-Concept:

"authentication AND security NOT testing"
→ Finds security-related auth discussions, excluding test-related messages

1. Click the search icon (🔍) in the top toolbar or press Cmd/Ctrl + K
2. Type your query in natural language
3. View results as you type (debounced after 300ms)
4. Click a result to jump to that message in context

Search Bar Features:

┌─────────────────────────────────────────┐
│ 🔍 Search messages with AI...         │
│                                         │
│ ✨ Semantic search enabled             │
└─────────────────────────────────────────┘

Available Filters:

Combining Filters:

authentication in:#engineering from:sarah after:2026-01-15
→ Sarah's messages about authentication in #engineering since Jan 15

Each search result shows:

┌─────────────────────────────────────────┐
│ 💬 Sarah in #engineering               │
│ 15 minutes ago                          │
│                                         │
│ "We fixed the authentication bug by... │
│  ...implementing proper session..."    │
│                                         │
│ 🎯 Relevance: 95% | Type: Semantic     │
│ 📝 2 replies                            │
└─────────────────────────────────────────┘

Elements explained:

• Relevance Score: How well the message matches your query (0-100%)
• Match Type:
  • Semantic - AI understanding (best)
  • Keyword - Word matching (fallback)
  • Exact - Exact phrase match (highest precision)
• Highlights: Relevant excerpts from the message
• Context: Optional before/after messages for context

To see surrounding messages:

1. Hover over a search result
2. Click "Show Context"
3. View 2 messages before and after (configurable)

Why context matters: Understanding the conversation flow around a message helps you grasp the full meaning.

Choose how results are ordered:

To change ranking: Click the ⚙️ Settings icon in search → Select ranking strategy

✅ Do:

• Use natural language: "show me", "find", "what did"
• Be specific: "React performance optimization" vs "React"
• Use synonyms: Try different terms if first search doesn't work
• Combine filters: Narrow down with channel, date, user filters
• Use quotes for exact phrases: "critical production bug"

❌ Avoid:

• Single-letter searches: "a", "I" (minimum 2 characters)
• Too generic: "help", "question" (too many results)
• Overly complex queries: Keep it simple for best results

Find experts:

"who knows about GraphQL?"
→ People discussing GraphQL, their messages, expertise

Track decisions:

"approved OR decided OR agreed" in:#product-planning
→ Decision-making messages in product planning

Monitor bugs:

"bug OR issue OR error" has:thread after:yesterday
→ Recent bug reports with discussion threads

Save frequent searches:

1. Perform a search
2. Click ⭐ Save Search
3. Name your search (e.g., "Recent bugs in engineering")
4. Access from Saved Searches dropdown

Benefits:

• Quick access to common queries
• Shared saved searches (admin feature)
• Email alerts for saved search results (coming soon)

If you see "Keyword search" badge:

This means AI-powered semantic search is not available, but you still have full keyword search functionality.

Differences:

• ✅ Still works for exact matches and keywords
• ✅ Still supports all filters and operators
• ⚠️ Won't understand synonyms or concepts
• ⚠️ May return more irrelevant results

Example:

• With semantic: authentication issues finds login, password, session problems
• Without semantic: authentication issues finds only messages with those exact words

Bots are automated assistants that can:

• Respond to commands (/help, /poll, /remind)
• Answer questions based on chat history and knowledge
• Automate tasks (create polls, set reminders, fetch data)
• Monitor events (welcome new members, track reactions)

Admins can add custom bots from the Bot Marketplace or build their own using the Bot SDK.

Basic format: /command [arguments]

Examples:

/hello
→ "Hi there, Alex! 👋"

/hello Alice
→ "Hello, Alice! 🎉"

/joke
→ "Why do programmers prefer dark mode? Because light attracts bugs! 🐛"

Get command list:

/help
→ Shows all available commands from all enabled bots

Get bot-specific help:

/help poll
→ Shows all PollBot commands and usage

Get command details:

/help /poll
→ Shows detailed usage for /poll command

Mention a bot to get its attention:

@HelloBot what can you do?
→ "I can greet people and tell programming jokes! Try /hello or /joke"

Some AI-powered bots understand natural language:

@SearchBot find discussions about authentication
→ SearchBot performs a semantic search and returns results

@SummaryBot summarize today's messages
→ SummaryBot generates a channel digest for today

Step-by-step:

1. Create poll:

/poll "Should we have pizza or tacos for lunch?" "Pizza" "Tacos" --duration 30
→ PollBot creates poll with 30-minute duration

2. Vote:

/vote poll-123 1
→ Vote for option 1 (Pizza)

3. Check results:

/results poll-123
→ Shows current vote counts

4. Close poll:

/closepoll poll-123
→ Closes poll and shows final results

Quick reminder:

/remind in 30 minutes to check deployment
→ "⏰ Reminder set! I'll remind you in 30 minutes."

Scheduled reminder:

/remind tomorrow at 9am to review pull requests
→ "⏰ Reminder set for tomorrow at 9:00 AM"

List reminders:

/reminders
→ Shows all your active reminders

Cancel reminder:

/cancel-reminder 456
→ "Reminder #456 cancelled"

Set up welcome message (Admins only):

1. Enable welcome:

/setwelcome on
→ "Welcome messages enabled for this channel"

2. Customize message:

/welcomemessage "Welcome {user} to {channel}! Check out our guidelines in the pinned message."
→ "Welcome message updated"

3. Test it:

/testwelcome
→ Shows preview of welcome message

Placeholders:

• {user} - New member's name
• {channel} - Channel name
• {server} - Workspace name

Bots have limited permissions to protect your data:

What bots CAN do: ✅ Read messages they're mentioned in ✅ Send messages to channels they're in ✅ React to messages ✅ Access channel member lists ✅ Store bot-specific data

What bots CANNOT do: ❌ Read messages in channels they're not in ❌ Access private/DM messages (unless explicitly invited) ❌ Modify or delete user messages ❌ Access user credentials or personal data ❌ Make changes to workspace settings

As a User:

1. Go to Settings > Bots
2. View all available bots
3. See what bots are active in each channel

As an Admin:

1. Go to Admin Panel > Bots
2. Enable/disable bots workspace-wide
3. Configure bot permissions
4. View bot usage analytics
5. Add custom bots from marketplace

Auto-moderation uses AI to automatically detect and handle problematic content:

• Toxicity Detection: Identifies insults, threats, harassment
• Spam Detection: Catches repetitive messages, excessive links
• NSFW Detection: Flags inappropriate images (if enabled)
• Profanity Filtering: Blocks or masks offensive language

Detection Process:

1. Message sent → Auto-mod scans content
2. AI analysis → Checks for toxicity, spam, profanity
3. Score calculation → Generates risk score (0-100%)
4. Action determination → Based on threshold and settings
5. Action execution → Flag, hide, warn, or mute

Analysis Methods:

• TensorFlow.js models (runs locally, no data sent to external APIs)
• Pattern matching (rule-based detection)
• Behavioral analysis (message frequency, link counts)

What it detects:

• Identity attacks
• Insults and name-calling
• Threats
• Severe toxicity
• Obscene language

Example:

❌ Message: "You're an idiot and should be fired"
🚨 Detected: Toxicity (85%), Categories: Insult, Identity Attack
⚡ Action: Hidden, user warned

What it detects:

• Excessive capitalization
• Repetitive characters
• Too many links
• Shortened URLs (bit.ly, etc.)
• High message frequency
• Known spam phrases

Example:

❌ Message: "CLICK HERE NOW!!! bit.ly/sketchy-link FREE MONEY!!!"
🚨 Detected: Spam (92%), Reasons: Excessive caps, shortened URL, spam phrases
⚡ Action: Message hidden

What it detects (images only):

• Pornographic content
• Sexually suggestive images
• Graphic violence

Note: NSFW detection is opt-in and must be explicitly enabled by admins.

What it does:

• Detects offensive words
• Can block, mask, or just flag
• Supports custom word lists
• Context-aware (avoids false positives)

Modes:

• Block: Message rejected, not sent
• Mask: Offensive words replaced with ***
• Flag: Message sent but flagged for review

Configurable by admins - Default actions can be customized per workspace.

You'll see:

⚠️ Your message may violate community guidelines

Your message was flagged for: Profanity

Options:
[Edit Message] [Appeal] [Cancel]

What you can do:

1. Edit Message: Revise and resend
2. Appeal: Request human review
3. Cancel: Don't send the message

You'll see:

[Hidden message]
This message was hidden by auto-moderation.
Reason: Policy violation

[Show anyway (Admin only)]

Regular users: Cannot see hidden messages Moderators/Admins: Can view and restore if appropriate

Admins and moderators can review flagged content:

1. Go to Admin Panel > Moderation Queue
2. View all flagged messages
3. For each item:
   • Approve: Mark as false positive, restore message
   • Delete: Confirm violation, keep hidden
   • Warn User: Send warning without other action
   • Mute User: Temporary or permanent mute

If you believe your message was incorrectly flagged:

1. Click "Appeal" on the moderation notice
2. Provide context or explanation
3. Submit appeal
4. Moderator review: Typically within 24 hours
5. Notification: You'll be notified of the decision

Tips for successful appeals:

• Explain context (was it a quote, reference, etc.?)
• Be respectful and constructive
• Provide relevant information

AI moderation isn't perfect. Common false positives:

• Technical jargon: "Kill the process", "abort transaction"
• Quotes: Quoting someone else's problematic message
• Non-English: Some languages may be misinterpreted
• Sarcasm: AI may not understand tone

What happens:

• Appeals are reviewed by humans
• False positives are marked to improve the system
• Custom allowlists can be created (admin feature)

Admins can configure:

• Thresholds: Adjust sensitivity (stricter or more lenient)
• Actions: Choose automatic actions per risk level
• Custom word lists: Add blocked or allowed words
• Exemptions: Exclude specific channels or users
• Notifications: Configure who gets notified of violations

See: Admin Guide - AI Management for details

For AI features, we collect:

When you use OpenAI-powered features:

• Messages sent to OpenAI for processing
• OpenAI's data usage policy applies
• Data is NOT used to train OpenAI's models (per enterprise agreement)
• Data is processed and discarded

When you use Anthropic-powered features:

• Messages sent to Anthropic for processing
• Anthropic's data usage policy applies
• Data is NOT used for training
• Data is processed and discarded

When using local fallbacks:

• No data sent to external services
• All processing happens on your device/server
• Complete privacy, but reduced capability

You can: ✅ Opt out of AI features entirely ✅ Request data deletion (GDPR right to erasure) ✅ Export your AI interaction history ✅ See what data is being processed ✅ Choose between AI providers

You cannot: ❌ See other users' AI interactions ❌ Access raw AI model data ❌ Opt out of auto-moderation (policy enforcement)

nself-chat is designed for GDPR compliance:

• Right to access: View your AI data in Settings
• Right to erasure: Request deletion of AI-related data
• Right to portability: Export your data in JSON format
• Data minimization: Only necessary data processed
• Consent: Explicit opt-in for optional AI features

For EU users: Data is processed per GDPR Article 6(1)(f) for legitimate interests (security, functionality) and Article 6(1)(a) for optional features (with consent).

Data in transit:

• ✅ TLS 1.3 encryption
• ✅ Certificate pinning
• ✅ End-to-end encryption for DMs (optional)

Data at rest:

• ✅ Encrypted database storage
• ✅ Encrypted backups
• ✅ Access logging and auditing

API communication:

• ✅ Secure API keys
• ✅ Rate limiting
• ✅ Request signing

To manage AI features for yourself:

1. Go to Settings > Privacy & AI
2. Configure preferences:

AI Features Preferences

Message Summarization
◉ Enabled (recommended)
○ Disabled

Reason: Uses AI to help you catch up on conversations

---

Smart Search
◉ Semantic search (AI-powered)
○ Keyword search only
○ Disabled

Reason: Understands meaning for better results

---

Bot Interactions
☑ Allow bots to see my messages in channels
☑ Allow bots to send me direct messages
☐ Allow bots to see when I'm online

---

Auto-Moderation
☑ Protect me from toxic content (recommended)
☑ Filter profanity in messages I see
☐ Strict mode (hide more content)

---

Data Usage
☐ Send anonymized AI usage data to improve features
☑ Allow AI processing of my messages for features

3. Click "Save Preferences"

Effect:

• ❌ No AI-generated summaries
• ✅ Basic statistical summaries still available
• ✅ All other features work normally

You'll see: "Summarization disabled in your settings"

Effect:

• ❌ No semantic search
• ✅ Keyword search still works
• ✅ All filters and operators available

You'll see: "Using keyword search (semantic search disabled)"

Effect:

• ❌ Bots won't respond to your messages
• ❌ Can't use slash commands
• ✅ See bot messages from others
• ✅ All human interactions normal

Note: You CANNOT completely opt out of auto-moderation as it's a workspace policy, but you can adjust how aggressively it filters content YOU see.

Options:

• Normal: Standard filtering
• Relaxed: Less aggressive (more content visible)
• Strict: More aggressive (safer experience)

Admins can control AI features for everyone:

1. Go to Admin Panel > AI Configuration
2. Enable/disable features:
   • Message Summarization
   • Smart Search
   • Bot Platform
   • Auto-Moderation
3. Set defaults for new users
4. Require/allow user opt-out

See: Admin Guide - AI Management

Quick disable for a session:

1. Click the AI indicator in the top toolbar
2. Toggle "Pause AI Features"
3. AI features disabled for current session
4. Re-enable anytime or automatically after session ends

Use cases:

• Privacy-sensitive conversations
• Testing non-AI functionality
• Reducing distractions

Symptoms: Greyed-out AI buttons, "Not available" messages

Causes & Solutions:

1. Workspace hasn't configured AI

   • ✅ Contact your admin to set up AI providers
   • ✅ Use local fallback features in the meantime

2. You've opted out

   • ✅ Go to Settings > Privacy & AI
   • ✅ Enable desired features

3. AI quota exceeded

   • ✅ Wait for quota reset (usually monthly)
   • ✅ Contact admin about upgrading plan

4. Network issues

   • ✅ Check internet connection
   • ✅ Try refreshing the page
   • ✅ Check if AI provider is experiencing outages

Symptoms: Searches only match exact keywords

Causes & Solutions:

1. No OpenAI API key configured

   • ℹ️ Semantic search requires OpenAI embeddings
   • ✅ Contact admin or use keyword search

2. Search cache full

   • ✅ Clear cache: Settings > Advanced > Clear AI Cache
   • ✅ Admin can increase cache size

3. Query too short

   • ✅ Use at least 2-3 words
   • ✅ Be more specific

Symptoms: Unhelpful or generic summaries

Causes & Solutions:

1. Using local fallback (no AI)

   • ℹ️ Check AI status indicator
   • ✅ Ask admin to configure OpenAI/Anthropic

2. Not enough content

   • ✅ Need at least 5-10 messages for good summaries
   • ✅ Try different time range

3. Messages lack context

   • ✅ AI works best with clear, substantive messages
   • ℹ️ Short "ok", "thanks" messages don't summarize well

4. Wrong summary type

   • ✅ Try different type: Brief vs Detailed vs Bullets
   • ✅ Adjust summary length in settings

Symptoms: Slash commands don't work, bot doesn't reply

Causes & Solutions:

1. Bot disabled

   • ✅ Check if bot is enabled: Admin Panel > Bots
   • ✅ Try in different channel

2. Wrong command syntax

   • ✅ Type /help to see correct syntax
   • ✅ Check for typos in command

3. Bot not in channel

   • ✅ Invite bot: /invite @BotName
   • ✅ Admin may need to enable bot for channel

4. Rate limit hit

   • ⏰ Wait a few minutes and try again
   • ℹ️ Bots have rate limits to prevent abuse

Symptoms: Your message was hidden or you received a warning unfairly

Causes & Solutions:

1. False positive

   • ✅ Click "Appeal" and explain context
   • ✅ Moderator will review within 24 hours

2. Technical jargon misinterpreted

   • ✅ Rephrase using different terms
   • ✅ Report to admin to add to allowlist

3. Quote or reference

   • ✅ Use quote formatting: > quoted text
   • ✅ Add context: "As mentioned earlier..."

If issues persist:

1. Check the Status Page: [status.yourworkspace.com]
2. Consult documentation: This guide and FAQ
3. Contact support:
   • In-app: Help > Contact Support
   • Email: [email protected]
   • Community: community.nself.org

Include in your support request:

• Description of the issue
• Steps to reproduce
• Screenshots (if applicable)
• Browser/app version
• AI features you're using

Q: Do I need to pay extra for AI features? A: It depends on your workspace's plan. Check with your admin. Personal API keys can be used in some configurations.

Q: Which AI provider is better: OpenAI or Anthropic? A: OpenAI is recommended for full semantic search support. Anthropic is excellent for summarization but doesn't support embeddings yet.

Q: Can I use my own AI API keys? A: Yes, in self-hosted deployments. Cloud/managed deployments use workspace-level keys only.

Q: Are AI features available offline? A: Local fallbacks work offline (keyword search, basic summaries, moderation), but AI-powered features require internet connectivity.

Q: Does nself-chat train AI models on my messages? A: No. We don't train models, and our agreements with OpenAI/Anthropic explicitly prohibit using your data for training.

Q: Can admins see my private messages if AI is enabled? A: No. AI features respect the same privacy boundaries as regular features. Admins cannot see private messages unless they're participants.

Q: Is auto-moderation watching all my messages? A: Auto-moderation scans messages for policy violations (like spam, toxicity) but doesn't store or analyze content beyond that. Think of it like a spam filter.

Q: Where is my data processed? A: Depends on the provider. OpenAI (US), Anthropic (US), or locally (for fallbacks). GDPR-compliant processing for EU users.

Q: Why does summarization sometimes take a long time? A: Large message volumes (100+ messages) take longer to process (5-10 seconds). Caching helps speed up repeated requests.

Q: Can I customize which bots are available? A: Yes, if you're an admin. Go to Admin Panel > Bots to enable/disable specific bots.

Q: How accurate is auto-moderation? A: Very accurate (95%+ for toxicity, 90%+ for spam) but not perfect. False positives are rare and can be appealed.

Q: Can I create my own bots? A: Yes! See the Bot SDK Documentation for a complete guide.

Q: What happens if I hit the AI rate limit? A: You'll see a message like "Rate limit reached. Please try again in a few minutes." Features will temporarily fall back to local implementations.

• Bot SDK Complete Guide - Build your own bots
• Smart Search Guide - Advanced search techniques
• Auto-Moderation Guide - Configure moderation settings
• AI API Documentation - For developers
• Admin Guide - AI Management - Admin controls

📺 Coming Soon:

• Getting Started with AI Features (3 min)
• Mastering Smart Search (5 min)
• Creating Your First Bot (10 min)
• Understanding Auto-Moderation (4 min)

We're constantly improving AI features. Share your feedback:

• In-app feedback: Help > Send Feedback
• Community forum: community.nself.org/ai-features
• Feature requests: github.com/nself/nself-chat/issues

Last Updated: January 31, 2026 Version: v0.7.0 Next Review: February 28, 2026