René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"feat: add /tokens /context /cost commands to show token usage and context size","articleBody":"### Feature Request: Token Usage \u0026 Context Visibility\n\nIt would be very useful to have visibility into token consumption and context size directly from the CLI, similar to Claude Code's `/context` and `/cost` commands.\n\n## Motivation\n\nCurrently, Deep Code already **stores token usage data internally** in `sessions-index.json`. According to the docs, each session entry tracks:\n\u003e \"token 用量 and 活跃 token 数\"\n\nHowever, there is **no way to view this information from the CLI interface**. Users have no visibility into:\n- How many tokens they have consumed in the current session\n- How much of the context window is being used\n- What is consuming the most tokens (system prompt, conversation history, skills, MCP tools)\n\n## Proposed Commands\n\n### `/tokens` – Show current session token consumption\n\nDisplays the token usage breakdown for the current session:\n- Input tokens (this turn)\n- Output tokens (this turn)\n- Thinking tokens (if thinking mode is enabled)\n- Total tokens (entire session)\n- Estimated cost (based on model pricing)\n\n### `/context` – Show context window breakdown\n\nDisplays what is consuming the context window:\n- System prompt size (tokens)\n- Conversation history size (# of messages, tokens)\n- Skills loaded and their contribution\n- Active MCP tools and their descriptions size\n- Total context window usage (e.g., 15K / 128K / 1M)\n\n### `/cost` – Show cumulative session cost\n\nShows the estimated cost of the current session based on model pricing:\n- Per-turn cost breakdown\n- Total session cost\n- Thinking vs non-thinking token cost\n\n## Why This Matters\n\n1. **Cost awareness** – Helps users monitor spend when using paid API models (especially with reasoning/thinking tokens which add significant cost)\n2. **Debug context limits** – When approaching context window limits, users can identify the culprit (bloated history, too many skills, etc.)\n3. **Optimization feedback** – Allows users to measure the impact of prompt optimization, skill trimming, or MCP cleanup\n4. **Parity with alternatives** – Tools like Claude Code already offer `/context` and `/cost`; adding these would reduce friction for users migrating to or evaluating Deep Code\n\n## Implementation Notes\n\n- The data already exists in `~/.deepcode/projects/\u003cproject-code\u003e/sessions-index.json` which stores \"token 用量\" per session entry\n- For real-time context window tracking, the token counts could be computed from the messages being sent to the model API\n- A lightweight tokenizer or cached counts per message would suffice — no need for full tokenization on every request\n\n---\n\n*This is an open-ended feature request. Happy to help with implementation or testing.*","author":{"url":"https://github.com/andrewsruge2501-max","@type":"Person","name":"andrewsruge2501-max"},"datePublished":"2026-07-07T02:05:17.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":2},"url":"https://github.com/224/deepcode-cli/issues/224"}