@logsesh/core

Parser, normalization, search, export-safety, and cost-estimation library behind logsesh.

Use it when you want to build your own local workflows on top of Claude Code, Codex, or Gemini CLI session logs.

Install

npm install @logsesh/core

Requires Node.js >=22.

Use

import { runPipeline, sanitizeForExport } from "@logsesh/core";

for await (const { session, warnings } of runPipeline({ toolFilter: ["codex"] })) {
  for (const warning of warnings) console.warn(warning.message);
  if (!session) continue;

  const safeSession = sanitizeForExport(session);
  console.log(safeSession);
}

Main Exports

| export | purpose | | --- | --- | | runPipeline | discover and parse sessions | | parseFile | parse one known log file | | sanitizeForExport | remove raw paths and reasoning by default | | searchSession | search a normalized session | | StatsAggregator | aggregate session stats | | estimateSessionCost | add labeled pricing estimates with confidence | | runDoctor | inspect adapter, discovery, export defaults, and pricing data |

Published Data

import sessionSchema from "@logsesh/core/schemas/session.json" with { type: "json" };
import models from "@logsesh/core/pricing/models.json" with { type: "json" };

Schemas are available under @logsesh/core/schemas/*. Pricing data is available under @logsesh/core/pricing/*.

Cost estimates are exact only when a parsed session includes a known priced model. Historical and uncertain model rows are preserved with confidence metadata so callers can explain where estimates came from.

Safety Defaults

sanitizeForExport removes reasoning blocks and raw source paths unless you explicitly request them.

const safe = sanitizeForExport(session);
const withReasoning = sanitizeForExport(session, { includeReasoning: true });
const withRawPaths = sanitizeForExport(session, { rawPaths: true });

Compatibility

Normalized sessions use the logsesh.session.v1 schema version.

logsesh.stats.v1 gained additive fields in v0.2.0:

• dailyBurn — chronological per-day session/turn/token totals (no gap-filling)
• tokenBreakdown — per-category token sums with observed field-presence flags and observedSessionCount

Existing JSON fields and semantics are unchanged.

License

MIT