overcast-to-sqlite

Save listening history and feed/episode info from Overcast to a SQLite database. Try exploring your podcast listening habits with Datasette!

If you simply want a page showing your recently listened episodes, try out the sister project overcast-to-pages.

• How to install
• Commands
• Authentication
• Fetching and saving updates
• Extending and saving full feeds
• Downloading transcripts
• Downloading chapters
• Downloading audio files
• Generating HTML pages
• Running all commands
• Listening statistics
• Searching
• Database schema
• See also
• Development

How to install

Requires Python 3.13 or newer.

Run it once without installing:

$ uvx overcast-to-sqlite

Install it permanently if you want overcast-to-sqlite available for later commands:

$ uv tool install overcast-to-sqlite

Commands

Run overcast-to-sqlite --help for a full list of options.

Authentication

Run this command to login to Overcast (note: neither your password nor email are saved, only the auth cookie):

$ overcast-to-sqlite auth

This will create a file called auth.json in your current directory containing the required value. To save the file at a different path or filename, use the --auth=myauth.json option.

If you do not wish to save this information you can manually download the "All data" file from the Overcast account page and pass it into the save command as described below.

Fetching and saving updates

The save command retrieves all Overcast info and stores playlists, podcast feeds, and episodes in their respective tables with a primary key overcastId.

$ overcast-to-sqlite save

By default, this saves to overcast.db but this can be manually set.

$ overcast-to-sqlite save someother.db

By default, it will use the cookie from auth.json if present. You can point to a different location using -a:

$ overcast-to-sqlite save -a /path/to/auth.json

Alternately, you can skip authentication by passing in an OPML file you downloaded from Overcast:

$ overcast-to-sqlite save --load /path/to/overcast.opml

By default, the save command will save any OPML file it downloads adjacent to the database file in archive/overcast/. You can disable this behavior with --no-archive or -na.

For increased reporting verbosity, use the -v flag.

Extending and saving full feeds

The extend command that will download the XML files for all feeds you are subscribed to and extract tags and attributes. These are stored in separate tables feeds_extended and episodes_extended with primary keys xmlUrl and enclosureUrl respectively. (See points 4 and 5 below for more information.)

$ overcast-to-sqlite extend

Like the save command, this will attempt to archive feeds to archive/feeds/ by default. This can be disabled with --no-archive or -na.

It also supports the -v flag to print additional information.

There are a few caveats for this functionality:

1. The first time this is invoked will require downloading and parsing an XML file for each feed you are subscribed to. (Subsequent invocations only require this for new episodes loaded by save) Because this command may take a long time to run if you have many feeds, it is recommended to use the -v flag to observe progress.
2. This will increase the size of your database by approximately 2 MB per feed, so may result in a large file if you subscribe to many feeds.
3. Certain feeds may not load due to e.g. authentication, rate limiting, or other issues. These will be logged to the console and the feed will be skipped. Likewise, an episode may appear in your episodes table but not in the extended information if it is no longer available.
4. The _extended tables use URLs as their primary key. This may potentially lead to unjoinable / orphaned episodes if the enclosure URL (i.e. URL of the audio file) has changed since Overcast stored it.
5. There is no guarantee of which columns will be present in these tables aside from URL, title, and description. This command attempts to capture and normalize all XML tags contained in the feed so it is likely that many columns will be created and only a few rows will have values for uncommon tags/attributes.

Any suggestions for improving on these caveats are welcome, please open an issue!

Downloading transcripts

The transcripts command downloads available transcripts for episodes.

The save and extend commands MUST be run prior to this.

Episodes with a "podcast:transcript:url" value will be downloaded from that URL and the download's location will then be stored in "transcriptDownloadPath".

$ overcast-to-sqlite transcripts

By default this will save transcripts to archive/transcripts/<feed title>/<episode title>.

A different path can be set with the -p/--path flag.

It also supports the -v flag to print additional information.

There is also a -s flag to only download transcripts for starred episodes.

Downloading chapters

The chapters command downloads and stores available chapters for episodes. The save and extend commands MUST be run prior to this.

$ overcast-to-sqlite chapters

By default, chapters are archived to archive/ adjacent to the database file. A different path can be set with the -p/--path flag.

Downloading audio files

The audio command downloads the audio files for episodes in your Overcast history (played, queued, or starred). Only the save command MUST be run prior to this (extend is not required).

$ overcast-to-sqlite audio

By default this will save audio files to archive/audio/<feed title>/<episode title>.<ext>.

The download's location is then stored in the enclosureDownloadPath column of the episodes table, and re-running the command skips episodes that were already downloaded.

A different path can be set with the -p/--path flag.

There is also a -s flag to only download audio for starred episodes.

It also supports the -v flag to print additional information.

Note: audio files are large (often 50–200 MB per episode), so this command is intentionally not part of all and must be run explicitly.

Generating HTML pages

The html command generates static HTML pages for recently played, starred, and deleted episodes.

For databases created by older versions, run save or extend once before generating HTML to normalize existing enclosure URLs. Both commands remove query parameters and deduplicate extended episodes, preserving the complete metadata of an existing clean URL row when available. If every URL in a duplicate group has query parameters, the earliest stored row is retained. This cleanup runs even when extend has no feeds to download; opening the database and retrieving episodes do not perform URL cleanup.

$ overcast-to-sqlite html

This produces three files: overcast-played.html, overcast-starred.html, and overcast-deleted.html in the same directory as the database file.

A different output directory can be set with the -o/--output flag:

$ overcast-to-sqlite html -o /path/to/output/

The directory passed to --output is created if it does not already exist.

Running all commands

The all command runs save, extend, transcripts, and chapters sequentially in a single invocation:

$ overcast-to-sqlite all

It supports the same -a/--auth and -v/--verbose flags as save.

Listening statistics

The stats command shows a summary of your listening habits:

$ overcast-to-sqlite stats

This displays total episodes played, total listening time, starred episodes, subscribed/removed feeds, and top podcasts ranked by episode count and listening time.

Searching

The search command performs full-text search across episodes, feeds, and chapters. Run save and extend first, and run chapters as well if you want chapter results.

$ overcast-to-sqlite search "machine learning"

Results are grouped by category (episodes, feeds, chapters). Use --limit / -l to control the maximum results per category with a positive integer (default: 20).

$ overcast-to-sqlite search "interview" -l 5

Database schema

Core tables

Key columns

feeds: overcastId, title, subscribed, overcastAddedDate, notifications, xmlUrl, htmlUrl, dateRemoveDetected

episodes: overcastId, feedId, title, url, overcastUrl, played, progress (seconds), enclosureUrl, userUpdatedDate, userRecommendedDate (starred date), pubDate, userDeleted, enclosureDownloadPath (from audio)

feeds_extended: xmlUrl (FK to feeds), title, description, lastUpdated, link, guid, plus dynamic columns from RSS XML

episodes_extended: enclosureUrl (FK to episodes), feedXmlUrl (FK to feeds_extended), title, description, link, guid, plus dynamic columns from RSS XML

chapters: enclosureUrl (FK to episodes), guid, source, time (seconds), content, url, image

Views

Full-text search indexes

FTS5 indexes are available on:

• feeds_extended (title, description)
• episodes_extended (title, description)
• chapters (content)

These are queried by the search command, or directly via SQL with MATCH syntax.

See also

• Datasette
• Podcast Transcript Convert
• Overcast Parser
• Podcast Archiver

Development

Pull requests are very welcome! For major changes, please open an issue first to discuss what you would like to change.

Setup

git clone git@github.com:hbmartin/overcast-to-sqlite.git
cd overcast-to-sqlite
uv sync --dev
uv run overcast-to-sqlite all -v

Building distributions

Build the source distribution and wheel with the uv_build backend:

uv build

Code Formatting and Linting

uv run black overcast_to_sqlite
uv run ruff check overcast_to_sqlite --fix
uv run pyrefly check overcast_to_sqlite
uv run ty check overcast_to_sqlite
uv run pytest tests/

This project is linted and formatted with ruff. Type checking is done with pyrefly and ty.

Experimental garuff checks

The lint workflow also runs garuff in a separate, advisory job. It checks the application, tests, and project documentation using the configuration in pyproject.toml. Findings and checker failures do not block CI while the tool is being evaluated. Each run records the source commit, uv version, exit code, and full output in the job summary and uploads a garuff-report artifact retained for 14 days.

garuff currently requires Python 3.14 and has no packaged release. CI installs it in an isolated tool environment from a pinned Git commit, so the project's Python 3.13 support and development dependencies are unaffected. The garuff job also pins uv to 0.11.26, which satisfies garuff's uv_build>=0.11.26,<0.12.0 requirement and avoids the build-system version warning seen with uv 0.12.24. Run the same toolchain locally with:

uvx --from 'uv==0.11.26' uvx --python 3.14 --from 'git+https://github.com/gahjelle/garuff@7d126c00f13b4d94f39f1107934089241ab7a6a1' garuff check overcast_to_sqlite tests README.md AGENTS.md CLAUDE.md

The configuration disables GAC001 to preserve postponed annotations needed on Python 3.13, GAC005 to allow homogeneous tuples, GAC008 to preserve positional parameters used by inherited method interfaces, and GAC010 to allow functions without docstrings. All other rules remain enabled. CI runs without --fix.

The Playlist, Feed, and Episode model constructors require keyword arguments for every field. For example, use Playlist(title="Queue", smart=0, sorting="newest", includePodcastIds="1,2"). Positional model construction now raises TypeError.

The initial evaluation on 2026-10-10, against application revision 37fbd52, completed without checker crashes or parse errors and exited with status 1 for 216 findings: 91 in the application, 125 in tests, and none in the three Markdown files checked.

A follow-up run with the three rules disabled and uv pinned to 0.11.26 rebuilt garuff with an empty cache. It completed without the build-system version warning, checker crashes, or parse errors and exited with status 1 for 68 remaining findings: 41 in the application, 27 in tests, and none in the three Markdown files checked.

The latest run also disables GAC008, makes all three model dataclasses keyword-only, and uses single backticks in the affected docstring. An empty-cache build and check completed with 0 findings across 24 Python files and three Markdown files, exit status 0, and no build-system version warning, checker crashes, or parse errors.

The compatibility concerns were reproduced without changing application files: applying the GAC001 advice to feed.py caused NameError: name 'Path' is not defined on Python 3.13, and applying GAC008 to the HTML parser override caused a TypeError during parsing. Keep the source and uv pins and advisory status until upgrades have been evaluated.

Packaging metadata checks

pyroma621 checks the project's packaging metadata and build configuration. It is pinned to version 6.1.0 in the development dependencies while its stability is evaluated; the installed command is pyroma.

Run the same check used in CI:

uv run --locked pyroma --min 10 .

The target is 10/10. Exit code 0 means the target was met, 2 means the rating is below the target, and 3 means an error prevented the check. For structured findings, use uv run --locked pyroma --format json --min 10 .. Version 6.1.0 also reports advisory findings, which do not affect the rating unless --strict is used.

CI runs this experimental check on Python 3.13 and 3.14. Failures produce a workflow warning and remain advisory while the checker is evaluated. Review the check's log for metadata findings or checker errors.

Version 6.1.0 fixes the two issues observed in 6.0.0: malformed TOML now returns exit code 3 with a structured JSON error, and JSON reports include the checker version in _meta.pyroma. Treat any nonzero exit code as a failed check.

Authors

• Harold Martin - harold.martin at gmail