# Uips Log Parser

> Parse and analyse UiPath Studio local execution logs to diagnose automation runs. Use when a developer asks about errors, warnings, duration, or status of a UiPath process run; when troubleshooting a failed, interrupted, or slow automation; when searching for a specific message across multiple runs. Triggers on: UiPath log, execution log, Studio run, robot error, job failed, process crashed, how long did it run, find in log.

- Skill: `cprima-agents/uips-log-parser` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add cprima-agents/uips-log-parser`
- Raw SKILL.md: https://api.skillmd.com/api/skills/cprima-agents/uips-log-parser/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- License: Apache-2.0
- Author: cprima-agents (https://skillmd.com/u/cprima-agents)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/cprima-agents/uips-log-parser

---


# UiPath Studio Execution Log Parser

Parse local UiPath Studio execution logs grouped by job run (`jobId`).

## Run the script

```bash
# Last job (default)
uv run skills/uips-log-parser/scripts/parse_logs.py

# Specific file
uv run skills/uips-log-parser/scripts/parse_logs.py --file "C:/Users/<user>/Downloads/2026-03-25_Execution.log"

# Last N runs from auto-discovered logs
uv run skills/uips-log-parser/scripts/parse_logs.py --last 5

# All runs, errors only
uv run skills/uips-log-parser/scripts/parse_logs.py --all --errors-only

# Search for a string across all runs
uv run skills/uips-log-parser/scripts/parse_logs.py --all --needle "Object reference"

# Show all log entries including Trace/Verbose (framework internals)
uv run skills/uips-log-parser/scripts/parse_logs.py --min-level Trace

# Duration between two copied log lines
uv run skills/uips-log-parser/scripts/parse_logs.py --duration "LINE1" "LINE2"

# Structured output for further processing
uv run skills/uips-log-parser/scripts/parse_logs.py --last 3 --format json
```

All diagnostics go to stderr; `--format json` sends clean JSON to stdout.

## Understanding the output

Each run shows a status flag:

| Flag | Meaning |
|------|---------|
| `[ OK ]` | Execution ended normally, no errors |
| `[FAIL]` | Execution ended normally, with errors |
| `[ .. ]` | No "execution ended" message — may still be running, was interrupted, or log is truncated |

**Do not assume `[ .. ]` means failure.** The run may be active right now.

## Log auto-discovery

The script searches these locations by default:

- `%LOCALAPPDATA%\UiPath\Logs\` — Studio / attended Robot (user-level)
- `C:\ProgramData\UiPath\Logs\` — Robot service (system-level)

Use `--log-dir PATH` to override, or `--file PATH` to point at a single file.

## Key CLI flags

| Flag | Default | Purpose |
|------|---------|---------|
| `--last N` | 1 | Show last N runs |
| `--all` | off | Show all runs |
| `--days N` | 7 | Discovery lookback window |
| `--date YYYY-MM-DD` | — | Filter to one day |
| `--process NAME` | — | Filter by process name (substring) |
| `--needle TEXT` | — | Filter runs containing TEXT; highlights matches |
| `--errors-only` | off | Show only runs with errors |
| `--warnings` | off | Also show warning details |
| `--min-level LEVEL` | — | Show all entries ≥ LEVEL (Trace/Verbose/Information/…) in an "All entries" section |
| `--format json\|text` | text | Output format |
| `--list-files` | off | Show discovered files and exit |

## Windows encoding

UiPath log messages may contain Unicode characters (e.g. `→`) that crash the script when the Windows console uses its default `cp1252` encoding. Set one of these before running:

```powershell
$env:PYTHONUTF8 = "1"
```

or per-invocation:

```powershell
$env:PYTHONUTF8 = "1"; uv run skills/uips-log-parser/scripts/parse_logs.py --min-level Trace
```

## Gotchas

- Log files start with a UTF-8 BOM (`\xEF\xBB\xBF`) — the script strips it automatically.
- The `[ .. ]` status does **not** indicate a crash when the user is actively developing in Studio.
- Duration (`totalExecutionTime`) is only present on the final "execution ended" entry. If the run is open, duration will be `null` in JSON / `?` in text.
- Multiple runs per day are in the same file, separated only by `jobId`.
- The `execution_log_data/` subfolder contains a SQLite DB (cloud-synced data) — the script does not read it.