Obsidian Stats logoObsidian Stats

TokenBar

by sn1p4am
5
4
3
2
1
Score: 50/100
Install Code

Description

Reviews

No reviews yet.

Stats

1
stars
214
downloads
0
forks
68
days
68
days
68
days
0
total PRs
0
open PRs
0
closed PRs
0
merged PRs
0
total issues
0
open issues
0
closed issues
9
commits

Latest Version

0.1.3
2 months ago

Changelog

Fix community review source warnings while keeping local tokenizer assets under the Sync Standard size limit.

See all version on GitHub

README file from

Github

TokenBar

TokenBar is an Obsidian desktop plugin that shows the active note's local token count in the status bar.

Features

  • Status bar display: 1,234 tokens
  • Primary tokenizer setting, defaulting to GPT
  • Hover comparison for GPT, Qwen, DeepSeek, Claude (est.), and Gemini (est.)
  • Text-only note counting; embedded images and attachments are excluded

Usage

  1. Enable TokenBar from Obsidian's Community plugins settings.
  2. Open a Markdown note.
  3. Read the token count in the status bar.
  4. Hover over the TokenBar status item to compare all supported token counting methods.
  5. Open TokenBar's settings to choose the primary tokenizer shown in the status bar.

Tokenizers

  • GPT: js-tiktoken with o200k_base
  • Qwen: bundled Qwen3.5 tokenizer asset
  • DeepSeek: bundled DeepSeek V3.1 tokenizer asset
  • Claude and Gemini: lightweight local estimates

All counting is local. Claude and Gemini values are estimates and may differ from provider-side token counts.

Privacy

TokenBar runs fully locally. It does not send note content to remote services, does not include telemetry, and does not require an account or payment.

Limitations

  • Claude and Gemini counts are estimates because their provider-side tokenizers are not bundled.
  • Embedded images and attachments are excluded; TokenBar counts note text only.
  • TokenBar is desktop-only.

Third-party notices

TokenBar uses these tokenizer libraries and assets:

  • js-tiktoken (MIT)
  • tokenx (MIT)
  • @cyberlangke/tokkit-core (MIT)
  • @cyberlangke/tokkit-deepseek (MIT), with tokenizer assets derived from official Hugging Face repositories under the deepseek-ai organization
  • @cyberlangke/tokkit-qwen (Apache-2.0), with tokenizer assets derived from official Hugging Face repositories under the Qwen organization

Development

pnpm install
pnpm run build

For manual installation, copy main.js, manifest.json, and styles.css into:

YourVault/.obsidian/plugins/tokenbar/
Sponsors