Understand tamag0 — from your first thread to company-wide collaboration.
Browse the complete product documentation for people, generated from the same source used by AI assistants.
Discover tamag0
Overview
What tamag0 is, positioning, and the team model (company → humans → companions).
Read guide →Companions
Named, specialized agents that grow with each person — golden rules, best practices, corrections, and behavioral continuity; onboarding aligns the companion with its human's work, standards, and mandate before it joins the team.
Read guide →Desktop app
Full interface tour — Conversations/Activity/Thoughts tabs, the execution bar under the composer with Auto, provider, model, effort, size, and permission controls, eco mode, workspaces, reading filters, rendered markdown and typeset mathematical notation, close-to-tray (Windows/Linux), settings — macOS, Windows, Linux.
Read guide →Observation dashboard
The read-only browser view of your companions — Home, Explorer, and per-companion Activity, scoped to your own company; how invitation, activation, and sign-in work.
Read guide →Thread lifecycle
The durable graph of parent, child, and blocking threads; explicit human, companion, thread, and external wait states; automatic dependency resumption; escalation and resolution.
Read guide →FAQ
Common buyer questions — differences with other AI tools, onboarding, data locality, and model provider choices.
Read guide →Core capabilities
Memory
Retrieval over an evolving, typed, write-back memory — not a frozen document index — with meaning-based recall and exact-language boosting, context-aware domain recall, private defaults and deliberate company sharing, persistent across sessions and context limits.
Read guide →Continuous learning
Companions get durably better between sessions through daytime reflection, overnight consolidation, bounded self-improvement, and autonomous study; they reconnect new work with older knowledge, and the learning survives a model swap.
Read guide →Collaboration
Real-time companion dialog, peer reviews, routing and human escalation; the distinction between persistent company companions and temporary runtime subagents.
Read guide →Cross-workspace missions
Place an experienced companion on a revocable assignment in another workspace — its accumulated knowledge is available by default, internal rules kept in the home workspace before the assignment stay behind, and client work remains excluded from home-workspace memory and overnight learning.
Read guide →Skills
Reusable workflows (SKILL.md convention), private or shared company-wide.
Read guide →Reflexes
The behavioral layer — golden rules, best practices, forbidden commands, and reminders — in two tiers (prompt-level judgment vs. hook-level enforcement), and when to reach for each.
Read guide →Library
One human-scoped file library reachable from every companion — streamed uploads via companion or native picker, deterministic folders, best-effort text extraction with processing/degraded states, private or company-wide visibility.
Read guide →Scheduled tasks and watchdog
Recurring routines, registered external signals (CI, PRs, timers), bounded companion wakes and human reminders, dependency rescue, auditability, and delegated task backlogs.
Read guide →Integrations
Slack, Gmail, Google Calendar, Jira, GitHub, Sentry, business-email intelligence — extensible through MCP.
Read guide →Models & operations
Architecture
What runs locally (app, agent runtime, one working directory per thread, OS-keychain secrets) vs on the shared company platform (memory, identity, threads, dialog hub) — what carries over across machines and what is recreated; the per-thread clone model (no single checkout, no git worktrees).
Read guide →Model providers
Claude, Codex, Ollama, or any OpenAI-compatible endpoint — reuses your existing Claude and ChatGPT subscriptions (no separate API billing), per-agent choice, automatic failover, fully local option.
Read guide →Claude Code CLI
The same companion, memory, and skills in any terminal session.
Read guide →Security
Per-company isolation, OS-keychain secrets, human consent on sensitive actions, audit trail.
Read guide →Performance
Built-in context compression (20% fewer tokens for coding agents, 60–95% fewer tokens for JSON, same answers), right-sized execution.
Read guide →Build with tamag0
AI-readable documentation
Use the compact index for discovery or the complete reference when you need every guide in one file.