TokenTracker

What the company appears to be: TokenTracker is a local-first tool for tracking token usage across 27 AI coding tools. The author states it works entirely on the user’s machine, with no cloud account, API keys or setup required. It parses session logs from AI tools locally and displays normalized token data in a dashboard that can be served locally or embedded in native desktop apps (macOS and Windows). An optional cloud layer enables cross-device aggregation and social features like leaderboards and badges.

What changed: The project started as a personal solution to a problem the author had with tracking usage across multiple AI coding tools. It evolved into a self-contained, local-first product with support for multiple platforms and an optional cloud sync feature.

The single most important open question: Is there any evidence of user adoption or traction beyond the author’s own use? The description states no revenue, customers or adoption data are available.

• The description states that TokenTracker is a local-first tool for tracking token usage across 27 AI coding tools.
• It parses session logs from these tools locally and normalizes them into consistent shapes (input tokens, cached input, cache-creation, output, reasoning tokens).
• The parsed data is served through a small local HTTP API on port 7680.
• A React + TypeScript dashboard is built once and reused across multiple surfaces: as a locally-served page, a Vercel-hosted marketing site, a macOS menu-bar app (Swift), and a Windows system-tray app (.NET + WebView2).
• The author states that an optional cloud layer supports cross-device aggregation and social features like leaderboards and badges.
• No cloud account, API keys or setup are required for local usage.

Note: The product is described as working entirely on the user’s machine, with no data leaving the device unless explicitly opted in via a cloud sync feature. The dashboard can be embedded in native apps but does not appear to require any third-party integrations beyond its own CLI and local parsing logic.

• The author states that TokenTracker was built because existing community tools failed to match provider billing data due to double-counting issues.
• It is positioned as a solution for users who distrust the accuracy of other tools and want privacy — no prompts or file contents are sent to any server.
• The tool evolved from a personal need into a product with cross-platform support and optional cloud features.
• The author emphasizes that the design principle was "local-first" from day one, which became a constraint rather than a feature.

Inference: The positioning appears to be driven by a combination of accuracy concerns and privacy preferences. There is no evidence of a broader market positioning or branding strategy beyond this personal motivation.

• The description states that the tool targets users who use multiple AI coding tools (e.g., Claude Code, Codex, Cursor, Gemini CLI).
• It is aimed at developers or technical users who care about tracking token usage and cost across these tools.
• Users are likely those who already use several AI coding tools and want to understand their combined usage without relying on fragmented dashboards.

Note: No evidence of customer segmentation, personas, or specific buyer profiles beyond the author’s own use case is provided. The target audience seems to be self-defined rather than market-tested.

• The description states that no cloud account, API keys or setup are required for local usage.
• An optional cloud layer exists for cross-device aggregation, leaderboards, and badges.
• No pricing information, monetization strategy, or revenue model is stated in the description.

Inference: There may be a freemium or tiered model where basic tracking is free and premium features (e.g., cloud sync) are paid. However, this remains unconfirmed.

• Built with JavaScript, plpgsql, Swift, TypeScript.
• Uses a CLI (bin/tracker.js) to install lightweight hooks into session logs on disk.
• Parses logs incrementally and normalizes schemas from different providers.
• Data lands in an append-only local queue and is served via a local HTTP API on port 7680.
• Dashboard built with React + TypeScript, reused across platforms (local web, macOS app, Windows app).
• Cross-platform support includes native macOS menu-bar app and Windows system-tray app.
• Cloud sync is additive — only aggregate token counts are shared, not prompt content.

Note: The tool is described as self-contained and not requiring external dependencies beyond its own runtime. Packaging challenges were noted for both macOS and Windows apps.

• No evidence of revenue, customers, or adoption beyond the author’s personal use.
• The project was submitted to the OpenAI 2026 hackathon on Devpost.
• There is no mention of user feedback, usage metrics, or product iteration history beyond the author's own account.

Absence of evidence: No data on traction, growth, or user engagement is provided. The maturity level is inferred from the complexity of the solution but not confirmed.

• The description does not name specific competitors.
• It mentions that existing community tools failed to match provider billing data due to double-counting issues.
• The author notes that no vendor-published spec exists for token count sums, making validation difficult.

Inference: TokenTracker appears to address a gap in accuracy and privacy among existing tools. However, the competitive landscape is not clearly defined or described.

• The tool is described as being built by a single person (team size: 1).
• No evidence of product-market fit, user feedback, or traction.
• The author notes that packaging for macOS and Windows was difficult — suggesting potential scalability issues.
• Cross-device aggregation introduces design complexity and security risks (e.g., missing row-level-security policy).
• The tool is local-first but includes an optional cloud layer — this tension may complicate long-term maintenance.

Inference: The lack of external validation, user data, or team size raises questions about scalability and commercial viability. The optional cloud feature introduces potential security and design trade-offs.

1. What is your evidence that users care enough to pay for this tool?
2. How do you plan to scale beyond a single developer’s use case?
3. Have you validated the accuracy of your parsers against real provider billing data across all 27 tools?
4. Is there any feedback or interest from users outside of yourself?
5. What is the roadmap for monetization and product development?
6. How do you plan to handle security and privacy concerns as more users adopt the cloud sync feature?