codexbar cost

Synopsis

codexbar cost [OPTIONS]
codexbar cost --format json --pretty
codexbar cost --provider claude --refresh

The cost command analyzes local conversation history to calculate token usage and estimated costs for Claude and Codex. It does not require web or CLI access, making it useful for offline cost tracking.

Parameters

--provider

string

default:"enabled providers in config"

Provider ID to analyze. Options:

Specific provider ID (e.g., codex, claude)
both - Analyze both Codex and Claude
all - Analyze all providers with local conversation history

Provider IDs are defined in ~/.codexbar/config.json.

--format

string

default:"text"

Output format. Options:

text - Human-readable output with cost summaries
json - Machine-readable JSON output with detailed token breakdowns

--refresh

boolean

Ignore cached conversation scans and re-analyze all local data. Useful when conversation files have been updated.

--pretty

boolean

Pretty-print JSON output with indentation.

--json-output

boolean

Emit machine-readable JSONL logs on stderr.

--json-only

boolean

Suppress non-JSON output; errors become JSON payloads.

-v, --verbose

boolean

Enable verbose logging.

--log-level

string

Set log level: trace, verbose, debug, info, warning, error, or critical.

--no-color

boolean

Disable ANSI colors in text output.

-h, --help

boolean

Show help message.

-V, --version

boolean

Show version information.

Response Fields

JSON Output Structure

The cost command returns an array of provider cost payloads:

provider

string

required

Provider ID (e.g., codex, claude).

source

string

required

Data source: always local for cost command.

updatedAt

string

required

ISO 8601 timestamp of when cost analysis was performed.

sessionTokens

integer

Total tokens used in current session.

sessionCostUSD

number

Estimated cost in USD for current session.

last30DaysTokens

integer

Total tokens used in the last 30 days.

last30DaysCostUSD

number

Estimated cost in USD for the last 30 days.

daily

array

Array of daily token usage and cost breakdowns.

Show daily[] fields

date

string

required

Date in YYYY-MM-DD format.

inputTokens

integer

Input (prompt) tokens consumed.

outputTokens

integer

Output (completion) tokens generated.

cacheReadTokens

integer

Cache read tokens (prompt caching).

cacheCreationTokens

integer

Cache creation tokens (prompt caching).

totalTokens

integer

Total tokens for the day.

totalCost

number

Estimated cost in USD for the day.

modelsUsed

array

Array of model names used on that day.

modelBreakdowns

array

Per-model cost breakdown.

modelName

string

Model identifier (e.g., claude-3-5-sonnet-20241022, gpt-4o).

cost

number

Cost in USD for this model on this day.

totals

object

Aggregate token and cost totals.

Show totals fields

inputTokens

integer

Total input tokens across all days.

outputTokens

integer

Total output tokens across all days.

cacheReadTokens

integer

Total cache read tokens.

cacheCreationTokens

integer

Total cache creation tokens.

totalTokens

integer

Grand total of all tokens.

totalCost

number

Estimated total cost in USD.

Examples

Default cost analysis (text format)

codexbar cost

Output:

== Claude Local Cost Analysis ==
Session: 45,234 tokens ($0.23)
Last 30 days: 892,156 tokens ($4.51)

Daily breakdown:
  2025-12-04: 45,234 tokens ($0.23)
  2025-12-03: 78,901 tokens ($0.41)
  2025-12-02: 123,456 tokens ($0.68)
  ...

Models used: claude-3-5-sonnet-20241022, claude-3-opus-20240229

== Codex Local Cost Analysis ==
Session: 32,189 tokens ($0.18)
Last 30 days: 654,321 tokens ($3.12)

Daily breakdown:
  2025-12-04: 32,189 tokens ($0.18)
  2025-12-03: 56,789 tokens ($0.29)
  2025-12-02: 98,765 tokens ($0.51)
  ...

Models used: gpt-4o, gpt-4-turbo

JSON output with pretty printing

codexbar cost --format json --pretty

Output:

[
  {
    "provider": "claude",
    "source": "local",
    "updatedAt": "2025-12-04T18:30:15Z",
    "sessionTokens": 45234,
    "sessionCostUSD": 0.23,
    "last30DaysTokens": 892156,
    "last30DaysCostUSD": 4.51,
    "daily": [
      {
        "date": "2025-12-04",
        "inputTokens": 25100,
        "outputTokens": 18134,
        "cacheReadTokens": 1500,
        "cacheCreationTokens": 500,
        "totalTokens": 45234,
        "totalCost": 0.23,
        "modelsUsed": ["claude-3-5-sonnet-20241022"],
        "modelBreakdowns": [
          {
            "modelName": "claude-3-5-sonnet-20241022",
            "cost": 0.23
          }
        ]
      },
      {
        "date": "2025-12-03",
        "inputTokens": 44500,
        "outputTokens": 32401,
        "cacheReadTokens": 1800,
        "cacheCreationTokens": 200,
        "totalTokens": 78901,
        "totalCost": 0.41,
        "modelsUsed": ["claude-3-5-sonnet-20241022", "claude-3-opus-20240229"],
        "modelBreakdowns": [
          {
            "modelName": "claude-3-5-sonnet-20241022",
            "cost": 0.35
          },
          {
            "modelName": "claude-3-opus-20240229",
            "cost": 0.06
          }
        ]
      }
    ],
    "totals": {
      "inputTokens": 498234,
      "outputTokens": 385922,
      "cacheReadTokens": 6500,
      "cacheCreationTokens": 1500,
      "totalTokens": 892156,
      "totalCost": 4.51
    }
  },
  {
    "provider": "codex",
    "source": "local",
    "updatedAt": "2025-12-04T18:30:15Z",
    "sessionTokens": 32189,
    "sessionCostUSD": 0.18,
    "last30DaysTokens": 654321,
    "last30DaysCostUSD": 3.12,
    "daily": [
      {
        "date": "2025-12-04",
        "inputTokens": 18000,
        "outputTokens": 14189,
        "cacheReadTokens": 0,
        "cacheCreationTokens": 0,
        "totalTokens": 32189,
        "totalCost": 0.18,
        "modelsUsed": ["gpt-4o"],
        "modelBreakdowns": [
          {
            "modelName": "gpt-4o",
            "cost": 0.18
          }
        ]
      }
    ],
    "totals": {
      "inputTokens": 378456,
      "outputTokens": 275865,
      "cacheReadTokens": 0,
      "cacheCreationTokens": 0,
      "totalTokens": 654321,
      "totalCost": 3.12
    }
  }
]

Query specific provider

codexbar cost --provider claude --format json --pretty

Force refresh (ignore cache)

codexbar cost --refresh

Machine-readable output only

codexbar cost --json-only --format json

Exit Codes

0: Success
1: Unexpected failure
2: Provider binary not found (not applicable for cost command)
3: Parse/format error
4: CLI timeout (not applicable for cost command)

Notes

Data Sources

The cost command analyzes local conversation history stored by the menubar app
Does not require web scraping, CLI tools, or API access
Useful for offline cost tracking and budgeting
Claude: Analyzes conversations from ~/Library/Application Support/Claude/
Codex: Analyzes conversations from OpenAI API usage logs

Cost Estimation

Costs are estimates based on published pricing and may not reflect actual charges
Pricing varies by model:
- Input tokens, output tokens, and cache tokens are priced differently
- Some models have tiered pricing or volume discounts
Cache tokens (prompt caching) are typically cheaper than regular input tokens
Actual costs depend on your provider plan and any custom pricing agreements

Token Counting

Token counts are extracted from conversation metadata when available
For older conversations without metadata, tokens may be estimated using tokenizer libraries
Cache read/creation tokens are only counted for providers that support prompt caching (e.g., Claude)

Caching

By default, conversation scans are cached for performance
Use --refresh to ignore the cache and re-scan all conversations
Cache is automatically invalidated when new conversations are added

Supported Providers

Claude: Full support with prompt caching token breakdown
Codex (OpenAI): Full support for GPT-4, GPT-4o, GPT-3.5-turbo, etc.
Other providers may have limited or no local cost tracking

Time Windows

Session: Current session’s token usage (resets with the provider’s session window)
Last 30 days: Rolling 30-day window of token usage
Daily breakdown includes all days with recorded usage in the last 30 days

CLI Commands

Provider API

Synopsis

Parameters

Response Fields

JSON Output Structure

Examples

Default cost analysis (text format)

JSON output with pretty printing

Query specific provider

Force refresh (ignore cache)

Machine-readable output only

Exit Codes

Notes

Data Sources

Cost Estimation

Token Counting

Caching

Supported Providers

Time Windows

Build docs developers (and LLMs) love

CLI Commands

Provider API

​Synopsis

​Parameters

​Response Fields

​JSON Output Structure

​Examples

​Default cost analysis (text format)

​JSON output with pretty printing

​Query specific provider

​Force refresh (ignore cache)

​Machine-readable output only

​Exit Codes

​Notes

​Data Sources

​Cost Estimation

​Token Counting

​Caching

​Supported Providers

​Time Windows

Build docs developers (and LLMs) love

Synopsis

Parameters

Response Fields

JSON Output Structure

Examples

Default cost analysis (text format)

JSON output with pretty printing

Query specific provider

Force refresh (ignore cache)

Machine-readable output only

Exit Codes

Notes

Data Sources

Cost Estimation

Token Counting

Caching

Supported Providers

Time Windows