Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.teeem-ai.com/llms.txt

Use this file to discover all available pages before exploring further.

​
Coming soon
Next release
macOS and Windows native desktop appsNative apps with the same permissions, audit trail, and PII protections as web chat. System tray presence, global shortcut (Cmd+Shift+T / Ctrl+Shift+T), push notifications, offline cache.Beta access via sales.
​
2026-05
Latest release
Knowledge Base — document lifecycle v2
  • Three-tier dedup (byte hash → normalised fingerprint → vector similarity 0.92 gate → Gemini verification)
  • Re-uploaded document versions soft-delete the previous version and record a supersession reason
  • Version history panel in the admin console
  • Slack attachments flow into the KB processing queue automatically (/upload-on toggles per channel)
Agent Builder v1
  • Author custom agents, enable tools, and test live at /admin/agents
  • Saved definitions activate immediately — no redeploy
Channels
  • Microsoft Teams generally available
  • KakaoTalk bridge (Korean auto-replies, group-room mirroring to admin)
  • SMS (LMS/MMS) — 1-on-1 alerts and two-way conversation
Localisation
  • English and Japanese added (intake forms, chat, email notifications)
  • Default language per organisation; users can switch mid-conversation
Context efficiency
  • Older tool outputs auto-compressed after 3 turns
  • Automatic compaction at 70% context window
  • LLM-side prompt cache for system prompts → ~75% input cost reduction
​
2026-04
Confidence gate + proactive skills
Confidence gate
  • When answer confidence is low, the agent asks a clarifying question instead of guessing
  • On by default for every active organisation
Proactive skill generation
  • Every Monday at 09:00 KST the system detects repeated employee work patterns and proposes new skills
  • Slack DM to admins with [Approve] / [Reject] buttons
  • 90-day cooldown + automatic semantic dedup
File processing
  • Safe ZIP extraction (depth limit, ratio guard)
  • Vertex service-account auth for Gemini Files API multimodal calls
  • Files larger than 8 MB process asynchronously with chat-side completion push
​
2026-03
Vertex Gemini 3 standardisation
LLM provider consolidation
  • All LLM traffic standardised on Google Cloud Vertex AI Gemini
  • Simple / medium queries: Gemini 3.1 Flash Lite
  • Complex work: Gemini 3 Flash Preview
  • Conversation summarisation: Groq (helper)
Security
  • AI Basic Act Article 27 watermarking on by default
  • Korean prompt-injection detection patterns hardened
  • Compliance evidence package — exportable in one click

​
Want more detail?

Talk to sales

See how a release fits your organisation. Sales walks you through it.