tot-space

Your team sees the same picture. In real time.

tot-space is a shared development environment where everyone — developers, reviewers, leads, and AI agents — works in one space with shared context, real-time presence, and a single set of rules.

Team members, workspace map, and live feed

Shared workspace

One workspace. Everyone connected.

No more “which branch are you on?” or “did you see the agent output?” tot-space gives your team a shared, synchronized environment with live presence and context awareness.

Real-time presence

See who's online, what file they're editing, which agent sessions are running. Team awareness without context switching to Slack.

Synchronized views

Architecture, data models, and flow views update in real time for everyone. When an agent changes something, the whole team sees it immediately.

Shared agent sessions

Watch agent output as it happens. Comment on intermediate results. Pause, redirect, or cancel agent sessions collaboratively.

Co-editing

Edit together. Not over each other.

Live cursors and region locks let multiple developers work in the same file without conflicts. Built-in session chat keeps discussions next to the code they're about.

src/api/handlers.rs 3 editors · 1 agent
Alice Editing lines 14–38 · region locked
Bob Editing lines 92–104 · region locked
Agent-2 Writing tests in worktree-2 · isolated
Clara Reviewing — read-only view · commenting on line 26
Conflicts prevented by region locks · Agent changes merged after verification

Code review

Review through lenses, not raw diffs.

When an agent changes 12 files, a flat diff is useless. tot breaks the review into structural lenses — architecture, API contracts, data models, flows — so each reviewer checks what matters to them.

Scenario

Agent builds a new feature

Developer describes the task. Agent generates code across 12 files. tot runs every verification lens before anyone sees the result.

Developer Creates task “payment refund flow” → assigns to agent-1
Agent Generates 847 lines across 12 files in isolated worktree
tot Runs architecture, API, data model, flow, constitution, security checks
tot Constitution violation: rule #7 — payment idempotency missing
Agent Auto-fixes violation, re-runs checks → all pass
Reviewer Reviews via structural lenses → approves per architecture, API, data

Scenario

Team reviews together

Three people review the same agent session — each through their own lens. Discussions happen inline, not in Slack.

Developer Reviews code diff and flow logic → approves
Reviewer Checks architecture and API contracts → requests change in data model
Lead Verifies constitution compliance and security → approves
tot Merges approvals: 2/3 lenses approved, 1 pending data model fix
Agent Applies requested data model change → re-verified → all approved

Communication

Team chat inside the editor.

Channels per project, threads anchored to code, agent output shared inline. No more switching between Slack and your editor to discuss a change.

Project channels

Organized by project and topic. Messages are persistent, searchable, and linked to the codebase context at the time they were sent.

Code threads

Start a discussion from any line of code. Thread stays anchored to file:line:commit even as the code evolves. Create tasks or agent sessions from threads.

Agent integration

Agent results posted to channels automatically. Comment on agent output, request changes, or trigger a re-run — all from the chat.

Roles

Everyone has a role. Including AI.

tot-space splits work by responsibility, not by files. Each role gets its own set of tools, views, and permissions.

Developer

Writes code, launches agents, creates pipelines. Sees architecture and data model impact before changes land.

Reviewer

Reviews through structural lenses — not raw diffs. Approves per architecture, API, data models, and constitution.

Lead

Defines constitution, manages roadmap, sets agent budgets and forbidden actions. Controls what AI can and cannot do.

Scenario

Lead sets guardrails

Before any agent runs, the lead defines what’s allowed. Constitution rules, budgets, and forbidden actions apply to every session.

Lead Writes constitution: “no direct DB writes”, “all payments idempotent”
Lead Sets agent budget: $5/session, max 2 concurrent agents
Lead Forbidden actions: delete migrations, modify auth module, push to main
Agent Tries to modify auth → blocked by forbidden action rule
tot Logs violation, notifies lead, agent re-routes to allowed approach

Team constitution

Shared rules that everyone follows.

The team constitution is a single source of truth for architecture decisions, coding standards, and agent boundaries. Enforced automatically for both humans and AI.

For developers

Constitution rules are visible in the editor. Violations highlighted inline as you type. No guessing about project conventions — the rules are always there.

For agents

Constitution is injected into every agent prompt via RAG. After execution, LSP verification checks compliance. Violations block the change automatically.

For reviewers

Constitution status shown per lens in every review. If all rules pass, the reviewer focuses on logic and design — not on style or convention checks.

For leads

Define and evolve rules over time. See compliance trends across the team. Constitution audit log shows who changed what and when.

Activity

See what's happening across the team.

Activity timeline, daily stats, agent usage, and cost tracking. All in one place, updated in real time.

Activity timeline

Every commit, agent session, review, and discussion in chronological order. Filter by person, project, or action type.

Daily stats

Lines changed, agent sessions run, reviews completed, constitution violations caught. Per-developer and per-team breakdowns.

Cost tracking

AI spend per developer, per project, per agent. Budget utilization dashboards. Alerts when approaching limits.

Ready to bring your team into one workspace?

tot-space is in active development. Request early access or explore what it means for your business.