Tokonomics 6.0
AI context efficiency and usage visibility for Visual Studio Code.
Tokonomics helps reduce unnecessary context sent during AI-assisted development and
shows request-level token and cost information when reliable usage data is available.
Results vary by task, workspace, model, and provider.
Highlights
- Local-First Processing: Context-processing algorithms execute directly inside the local VS Code extension process without an external Tokonomics backend.
- Unified 14-Stage Context Pipeline: Intelligent multi-stage context selection, relevance scoring, deduplication, and budget fitting with deterministic fallbacks.
- Interactive Visual Dashboard: Real-time event-driven visualizer with live token stream charts, stage waterfalls, and request ledger.
- Multi-Language Support: Structural code pruning for TypeScript, JavaScript, Python, Go, Rust, Java, C, and C++.
- Workspace-Aware Chat Participant: Invoke
@tokonomics with dedicated commands like /dashboard, /explain, /map, and /pack.
- Privacy & Safety Safeguards: Zero-telemetry design, Restricted Mode compliance for untrusted workspaces, and independent component kill switches.
- Governed Preview Features: Optional experimental capabilities remain strictly off by default and require explicit consent.
Getting Started
Install the Tokonomics VSIX package or install it from your approved extension source.
Open the VS Code Chat view.
Enter a request using @tokonomics:
@tokonomics explain the authentication flow in this workspace
Open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P) and search for Tokonomics: Open Interactive Dashboard to view real-time metrics.
Tokonomics uses conservative defaults. Workspace-derived context is limited in
Restricted Mode and can be further controlled from VS Code settings.
Interactive Dashboard
The Tokonomics Dashboard (/dashboard or Tokonomics: Open Interactive Dashboard) provides:
- Dual Efficiency Streams: Live dynamic SVG charts visualizing token reduction and cost savings over time.
- Stage Waterfall Breakdown: Transparent visualization of context allocation across pipeline stages.
- Live Prompt Pulse Card: Instant state indicators (
OPTIMIZING ➔ OPTIMIZED ➔ RECONCILED) for active requests.
- Request Ledger & Inspector: Inspect decisions, retrieved context, and estimated vs. reconciled provider costs.
- Active File Diagnosis: One-click diagnosis and pruning simulation for the currently focused editor file.
- Multi-Window Aggregation: Switch seamlessly between Session, Today, 7-Day, and Lifetime metrics.
Chat Commands
| Command |
Purpose |
/dashboard |
Open the interactive activity dashboard with live visual charts |
/live |
View the current session token efficiency stream and cost metrics |
/explain |
Review the reasoning behind the latest context decision |
/stats |
View available usage totals and cache alignment statistics |
/map |
Generate a concise structural overview of the workspace |
/pack |
Prepare selected workspace files for AI context usage |
/analyze |
Run structural diagnosis on the active source file |
/compact |
Prepare and compress selected text for AI prompts |
/logs |
View sanitized local execution diagnostics |
/ram |
View configured local memory cache and index status |
Common Settings
Settings are available under Tokonomics in VS Code Preferences:
- Workspace Context Mode: Configure indexing depth and auto-context attachment.
- Local Memory Limit: Control memory ceilings for local caching and indexing.
- Response Cache Controls: Configure exact response caching and eviction policies.
- Release Channel & Emergency Disablement: Global kill switch to instantly bypass all optimization.
- Independent Kill Switches: Disable individual compiler components without restarting VS Code.
- Experimental Feature Consent: Explicit consent gate for preview features (all off by default).
Preview features are off by default. Enabling consent alone does not select a preview
feature, and every preview feature can be disabled independently.
Privacy and Security
Tokonomics performs its context preparation inside the local extension process and
does not require a Tokonomics-operated intermediary service. Prompts and selected
context are still sent to the AI provider chosen in VS Code.
The extension applies local safeguards and sanitized diagnostics, but users should
still review context before sending sensitive material. VS Code, other installed
extensions, and the selected AI provider have their own privacy and network behavior.
Compatibility and Validation
Tokonomics 6.0 requires VS Code 1.106.0 or later. Release builds are checked as exact
VSIX artifacts against the configured Windows minimum, Stable, and Insiders host
matrix. Other editors, operating systems, remote environments, provider accounts, and
billing behavior require separate qualification.
Repository benchmarks include controlled test fixtures. They validate expected
software behavior but do not guarantee production token savings, cost reduction, or
model-quality improvement.
Support
When reporting a problem, use the Tokonomics diagnostic export and review it before
sharing.
License
Tokonomics is proprietary software. See LICENSE.txt.
