# ICS Merge & Dedupe
Merge several .ics calendar files into one and remove duplicate events by UID or start+title — all in your browser. Free, private, nothing uploaded.
## Run it
- **CLI:** `gizza tool ics-merge-dedupe "BEGIN:VCALENDAR
VERSION:2.0
BEGIN:VEVENT
UID:evt-1@example.com
DTSTART:20240309T081530Z
SUMMARY:Team Standup
END:VEVENT
END:VCALENDAR
BEGIN:VCALENDAR
VERSION:2.0
BEGIN:VEVENT
UID:evt-1@example.com
DTSTART:20240309T081530Z
SUMMARY:Team Standup
END:VEVENT
END:VCALENDAR"`
- **Web:** https://gizza.ai/tools/ics-merge-dedupe/
- **Agents:** machine-readable descriptor (parameters JSON Schema) at https://gizza.ai/tools/ics-merge-dedupe/tool.json
## Inputs
- `ics` — .ics file contents (concatenate several to merge) _(field)_
- `dedupe_by` — Match duplicates by _(field)_
- `keep` — Keep which copy _(field)_
- `sort` — Sort events by start time _(field)_
- `calendar_name` — Calendar name (optional) _(field)_
## Output
- Merged calendar (text)
## Query parameters
Open the tool pre-filled and auto-run via URL:
- `ics` — .ics file contents (concatenate several to merge)
- `dedupe_by` — Match duplicates by
- `keep` — Keep which copy
- `sort` — Sort events by start time
- `calendar_name` — Calendar name (optional)
Example: `https://gizza.ai/tools/ics-merge-dedupe/?ics=BEGIN%3AVCALENDAR%0AVERSION%3A2.0%0ABEGIN%3AVEVENT%0AUID%3Aevt-1%40example.com%0ADTSTART%3A20240309T081530Z%0ASUMMARY%3ATeam%20Standup%0AEND%3AVEVENT%0AEND%3AVCALENDAR%0ABEGIN%3AVCALENDAR%0AVERSION%3A2.0%0ABEGIN%3AVEVENT%0AUID%3Aevt-1%40example.com%0ADTSTART%3A20240309T081530Z%0ASUMMARY%3ATeam%20Standup%0AEND%3AVEVENT%0AEND%3AVCALENDAR&dedupe_by=smart&keep=first&sort=true&calendar_name=Team%20Calendar`
---
## About this tool
**ICS Merge & Dedupe** combines one or more iCalendar files — the `.ics` format
that Google Calendar, Apple Calendar and Outlook export — into a single calendar
and removes duplicate events along the way. Paste the contents of one file, or
concatenate several files one after another, and you get back one clean
`BEGIN:VCALENDAR…END:VCALENDAR` document with each event kept only once.
Duplicates are matched the way real calendars overlap: by **UID** when both copies
carry one (the same event re-exported from the same source), or by **start time and
title** for the same public event that two different apps each stamped with their
own UID. Every surviving event is copied through **verbatim** — its properties,
folded lines, alarms and RFC 5545 escapes are untouched — so the result imports
cleanly back into any calendar app. Line endings are normalized to CRLF, identical
`VTIMEZONE` blocks are emitted once by their `TZID`, and `VTODO` / `VJOURNAL` /
`VFREEBUSY` components pass through unchanged.
Everything runs locally in your browser — your calendar never leaves your device.
### Worked example
Input — two files that both contain the same event (same `UID`):
```text
BEGIN:VCALENDAR
VERSION:2.0
BEGIN:VEVENT
UID:evt-1@example.com
DTSTART:20240309T081530Z
SUMMARY:Team Standup
END:VEVENT
END:VCALENDAR
BEGIN:VCALENDAR
VERSION:2.0
BEGIN:VEVENT
UID:evt-1@example.com
DTSTART:20240309T081530Z
SUMMARY:Team Standup
END:VEVENT
END:VCALENDAR
```
Output (defaults — smart matching, keep first, sorted) — one calendar, one event:
```text
BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//gizza-ai//ics-merge-dedupe//EN
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:evt-1@example.com
DTSTART:20240309T081530Z
SUMMARY:Team Standup
END:VEVENT
END:VCALENDAR
```
### Options
- **Match duplicates by** — `smart` (default) uses the UID when an event has one and
otherwise falls back to normalized start + title; `uid_start` matches UID *and*
start time, so recurrence overrides that share a UID but differ in start are kept;
`start_title` ignores the UID entirely and matches on normalized start + title,
collapsing the same event exported by different apps.
- **Keep which copy** — when a duplicate group is found, `first` keeps the earliest
occurrence in document order, while `last_modified` keeps the copy with the newest
`LAST-MODIFIED` (falling back to `DTSTAMP`) so a later edit wins.
- **Sort events by start time** — on by default; turn it off to keep events in the
order they appeared across the input files.
- **Calendar name** — optionally sets `X-WR-CALNAME`, the title Google/Apple/Outlook
show for the merged calendar.
### Limits & edge cases
- Only `VEVENT` blocks are deduplicated. `VTODO`, `VJOURNAL` and `VFREEBUSY` are
passed through unchanged (never merged), and at least one event is required.
- Title matching is whitespace-normalized and case-insensitive, so `Independence
Day` and `Independence Day` match — but genuinely different wording will not.
- Start times are compared after normalizing, so `20240309T081530Z` and its ISO
form compare equal. `Z` (UTC) times are exact; floating and `TZID` times are read
as wall-clock UTC because no timezone/DST database is shipped — the same wall-clock
value from two files still matches.
- Recurrence rules (`RRULE`) are **not** expanded — a repeating event stays a single
master event, and `uid_start` will not treat its instances as separate.
- Events with no UID, start *and* title at all carry no identity and are always
kept, never merged away.
## FAQ
How does it decide two events are the same?
It depends on the **Match duplicates by** option. With `smart` (the default), two
events match if they share a `UID`; an event with no UID falls back to matching on
its normalized start time and title. `uid_start` additionally requires the start
times to match, which keeps recurrence overrides that reuse one UID. `start_title`
ignores UIDs completely and matches on start + title, which is what collapses the
same public event that Google and Apple each exported under a different UID.
Which copy of a duplicate is kept?
With **Keep which copy** set to `first` (the default), the earliest occurrence in
the combined input wins and later copies are dropped. Set it to
`last_modified` to keep the copy with the newest `LAST-MODIFIED` timestamp
(falling back to `DTSTAMP` when that property is absent), which is what you want
when a second file is a more recently edited version of the first.
Are my events changed or reformatted?
No. Every surviving event is copied through **verbatim** — its properties, ordering,
folded continuation lines, `VALARM` sub-blocks and RFC 5545 text escapes are left
exactly as written. The only additions are a fresh `VCALENDAR` wrapper and, if you
provide one, an `X-WR-CALNAME` calendar-name line. Line endings are normalized to
CRLF so the output imports cleanly.
Is my calendar uploaded anywhere?
No. The merge runs entirely in your browser using WebAssembly — the `.ics` text you
paste is processed on your own device and is never sent to a server.
## Related tools
- [Preview a Bulk File Rename](https://gizza.ai/tools/bulk-file-renamer/): Preview old-to-new filename mappings with find/replace, regex, numbering, case conversion, prefix/suffix, and collision warnings.
- [Citation Generator](https://gizza.ai/tools/citation-generator/): Free APA 7, MLA 9, Chicago and Harvard citation generator — format author, title, year, journal and URL into a correct reference right in your browser.
- [Document Splitter](https://gizza.ai/tools/document-splitter/): Split a long Markdown or HTML document into separate files, one per top-level heading. Runs entirely in your browser, no upload, no sign-up.
- [ENEX to Markdown Converter](https://gizza.ai/tools/enex-to-markdown/): Convert Evernote ENEX exports to Markdown or plain text with titles, dates, tags, source URLs, and attachment summaries — locally in your browser.
- [File Tree Generator](https://gizza.ai/tools/file-tree-generator/): Turn a list of file paths or an indented outline into an ASCII tree for READMEs and docs. Unicode or ASCII connectors, optional sorting. Free, in-browser.