Skip to main content

Introducing Supporter Rate Limits

· 6 min read
Brian Pepple
Founder of the Metron Project / Code Monkey

Back when we added Open Collective support in December 2024, I mentioned that a paid tier for a higher API rate limit wasn't something I was planning on — but that it might be worth revisiting someday if there was enough demand. Then in March we had to cut the daily limit from 10,000 to 5,000 requests due to server load, with a promise to look at restoring it if donations picked up. Well, someday is here: donors to the project now automatically get an elevated daily API rate limit, no paid subscription plan required — just a donation through Open Collective.

June 2026 Updates

· 7 min read
Brian Pepple
Founder of the Metron Project / Code Monkey

June was a busy month for Metron. We shipped a major site UI refresh, a round of reading list improvements, several usability wins across issue and series pages, a new timeline page, account deletion support, and a handful of bug fixes — plus some news from third-party developers building on the Metron API. Here's a rundown of everything that landed.

Metron API Best Practices

· 9 min read
Brian Pepple
Founder of the Metron Project / Code Monkey

The Metron API gives developers programmatic access to a comprehensive comic book database — publishers, series, issues, characters, creators, story arcs, and more. To keep it fast and available for everyone, a little care in how you use it goes a long way. This post covers the patterns that will make your integration both efficient and a good citizen on the platform.

March 2026 News

· 9 min read
Brian Pepple
Founder of the Metron Project / Code Monkey

Monthly Statistics

During March the Metron Project added the following to its database:

  • Users: 164
  • Issues: 1,107
  • Creators: 36
  • Characters: 240

Thanks to everyone that contributed!

API Rate Limit Reduction

Early in March, we reduced the API rate limits due to increased server load from growing API usage. The new limits are:

  • 20 requests per minute (down from 30)
  • 5,000 requests per day (down from 10,000)

Updated releases of Mokkari and Metron-Tagger were published to reflect these changes. If you consume the API directly, make sure your code handles HTTP 429 responses and respects the Retry-After header. If we receive enough donations to cover a server upgrade, we would look at restoring the daily limit to 10,000 requests per day.

Website

Here's a summary of everything that landed in the website since the start of March, covering API improvements, reading list enhancements, collection stats, and a handful of performance and bug fixes.

API Improvements

Rate Limit Changes and Visibility

The API rate limit was adjusted to 20 requests/minute and 5,000 requests/day. To make this more transparent, the API now returns rate limit response headers on every response, so clients can see exactly how many requests they have remaining. Users can also see their current daily API usage — requests used, remaining, and a color-coded progress bar — directly in their profile's Account Settings section.

Conditional Request Support

Two more endpoints now support conditional requests (ETags / If-None-Match):

  • The reading list detail endpoint — item add/remove operations now bump the parent list's modified timestamp, so clients can detect item changes without polling /items/.
  • The /issue_list/ endpoints — thanks to Jonas Stendahl for this contribution.

This means clients can avoid re-downloading unchanged data, saving bandwidth for both users and the server.

Reading List Enhancements

The ReadingList model and API received several new additions:

  • list_type field — exposed in both the list and detail serializers, with human-readable display values.
  • image field — reading lists can now have a cover image. When set, the reading list detail page displays a semi-transparent hero header using the image; it falls back to the plain header when none is present.

Bug Fixes

  • Fixed a PATCH request to SeriesViewSet that was silently clearing the imprint field when it was omitted from the request body.
  • Fixed the ReadingList unique constraint to include attribution_source, allowing a user to have multiple lists with the same name as long as they come from different sources.

Reading List Web Views

  • Added a publisher filter to the reading list view, available in the Advanced Filters panel.
  • Fixed an issue count display bug in ReadingListListView.
  • Fixed a slow query in the publisher filter — the previous JOIN-based approach caused ~30 second response times on live data. We implemented a subquery-based approach is 5.6x faster.

Collection Stats

  • Added Top Writers and Top Artists horizontal bar charts to the Collection Stats view.
  • Fixed chart text readability in dark mode — chart tick labels, grid lines, legend text, and titles now follow Bulma's theme-aware colors rather than Chart.js's hardcoded defaults.

General UI / Frontend

  • Cover images in issue detail pages and card grids now handle landscape-oriented covers correctly. Instead of forcing all covers into a portrait 2by3 aspect ratio with a center crop, the template detects the aspect ratio of the generated thumbnail and uses is-3by2 for landscape images and is-2by3 for portrait.

Performance

  • Replaced prefetch_related with select_related in HomePageView for recently edited issues, reducing two extra queries to a single JOIN.
  • Eliminated unnecessary series type queries in several views — thanks to Jonas Stendahl for this contribution.
  • Fixed issue count annotations not appearing in PublisherSeriesList and ImprintSeriesList card grids.

Bug Fixes

  • Fixed an autocomplete widget crash (Internal Server Error) on create/update views when submitting a form without selecting an autocomplete field value. A SafeAutocompleteWidget subclass now normalises empty string and empty list values to None before the upstream widget tries to filter on them.

Dependency Updates

  • Updated to Django 5.2.12.
  • Updated Pillow.

Contributors

Thanks to Jonas Stendahl for his contributions this month.

Mokkari

Here's a summary of everything that landed in Mokkari during March, covering rate-limiting overhaul, new schema fields, and dependency refreshes.

Rate Limiting Changes

Reduced API limits

The per-session rate limits were updated to match the current Metron website limits:

LimitOldNew
Per minute3020
Per day10,0005,000

Pluggable rate-limit buckets

Session and the top-level api() function now accept an optional bucket parameter typed as AbstractBucket from pyrate_limiter. This lets you plug in any compatible backend — Redis, a database, or a plain in-memory store — to share rate-limit state across multiple workers or processes.

The default behaviour is unchanged: a SQLite-backed bucket is created lazily and shared across all sessions in the same process. A DEFAULT_RATES constant is now exported from mokkari.session so you can construct custom buckets with the same rate definitions:

from pyrate_limiter import RedisBucket
import redis
import mokkari

pool = redis.ConnectionPool.from_url("redis://localhost:6379")
bucket = RedisBucket.init(mokkari.session.DEFAULT_RATES, redis.Redis(connection_pool=pool), "mokkari")
m = mokkari.api("username", "password", bucket=bucket)

Contributors

Thanks to Myers Carpenter for his initial work on the pluggable rate-limit buckets this month.

Reading List Schema Additions

Two new fields were added to the reading list schemas:

  • image — an image URL field on ReadingListRead, surfacing the cover image for a reading list.
  • list_type — a new field on the reading list schemas that exposes the list's type classification.

Metron-Tagger

Here's a summary of everything that landed in Metron-Tagger during March, covering rate limit improvements, bug fixes, and dependency updates.

Rate Limit Handling

The tagger's handling of Metron API rate limits was improved:

  • Clock-skew buffer — A RATE_LIMIT_RETRY_BUFFER constant (2 seconds) is now added to every retry_after wait to account for small differences between the client clock and the server clock, reducing the chance of hitting the limit again immediately after waking up.
  • Nested rate limit handling — If a second rate limit error is encountered during a retry attempt, the tagger now performs one additional silent wait-and-retry instead of surfacing an error to the user. Only if that second retry also fails is an error reported.

These changes result in fewer spurious rate-limit errors when tagging large batches of comics (#238).

Bug Fixes

  • Fixed a crash that occurred when a MetronInfo file did not include an info_source field. Previously, iterating over md.info_source without checking for its existence would raise an exception. The fix adds an early None guard in get_id_from_metron_info so files without a source are silently skipped rather than crashing the tagger.

Dependency Updates

  • Updated the Mokkari dependency to pick up its latest improvements.

Contributors

Thanks to @bearinfo for his contributions this month.

Desaad

During March several new features around content discovery, reading list management, and library gap tracking were made.

Creator Views

Creators now have their own browsable pages. A new Creators section in the navigation gives access to a full creator list and individual detail pages showing a creator's roles across your library. A dedicated Creator Series/Issues view lets you drill down into their work by series.

Bookmarks

You can now bookmark series and issues directly from the list views. A bookmark icon overlays the cover image; clicking it toggles the bookmark state inline via HTMX without a page reload. A new Bookmarks page collects all your saved items in one place, with inline unbookmark support.

Advanced Search for Series and Issues

The Series and Issue list views gained advanced filtering capabilities. You can now filter by multiple criteria simultaneously, making it much easier to navigate large libraries.

Reading List Refresh with Conditional Requests

Reading lists sourced from Metron can now be refreshed. Desaad stores the modified datetime returned by Metron and uses it on subsequent refresh requests via If-Modified-Since. If Metron returns 304 Not Modified, you'll see an "already up to date" message. Otherwise, existing items are cleared and a background re-import is enqueued. A Refresh button appears on the detail page for any Metron-sourced list with a completed import.

Series Missing Issues Tracking

Desaad can now check your local library against Metron to identify issues in a series that you don't yet have. A new MissingSeriesIssue model tracks these gaps, and a Check Missing Issues button on the series detail page triggers the check inline via HTMX. Like reading list refresh, subsequent checks use conditional requests — if nothing has changed on Metron's side, you'll see an "already up to date" response.

Image of the missing issues on the Series Detail view

A dedicated Missing Issues list view provides an overview of all tracked gaps across your library, and a background task automatically cleans up missing issue records when you import a previously absent issue.

Reading List Improvements

Reading list search results and detail pages now show human-readable attribution source labels. Issue type display and filtering were added to reading list items, and attribution source tags appear on reading list cards for quicker at-a-glance identification.

Image of the reading list view

OpenCollective

A huge thank you to everyone who has contributed to our Open Collective! Your support makes a real difference in keeping the Metron Project running and growing.

What Your Contributions Support

Funds from Open Collective go directly toward:

  • Server hosting costs - Keeping the Metron website and API available
  • Domain registration - Annual domain name renewals
  • Future capacity increases - Scaling resources as the database and user base grows

All expenses are transparent and publicly viewable on our Open Collective page, so you can see exactly where every dollar goes.

Support the Project

If you'd like to help keep the lights on and support continued development, contributions of any size are appreciated and help ensure Metron remains a free resource for the comic book community.

Anyway, that's everything for this month! Take care.

API Rate Limit Reduction

· 2 min read
Brian Pepple
Founder of the Metron Project / Code Monkey

As mentioned in our February 2026 update, we have been dealing with increased server load due to the growth in API usage. We have made the decision to reduce the API rate limits effective immediately.

The new limits are:

  • 20 requests per minute (down from 30)
  • 5,000 requests per day (down from 10,000)

We know this is not ideal, and we're sorry for any inconvenience this may cause. If we receive enough donations in the future to cover the cost of a server upgrade, we would look at restoring the daily limit back to 10,000 requests per day.

Updated Client Libraries

New releases of our client libraries have been published that reflect these updated rate limits:

  • Mokkari v3.20.0 — the Python wrapper for the Metron API has been updated with the new rate limit configuration.
  • Metron-Tagger v4.9.0 — updated to use the latest version of Mokkari, picking up the new rate limits automatically.

If you use either of these tools, we recommend upgrading to the latest version.

Note for API Consumers

If you have written software that directly consumes the Metron API, please ensure your code correctly handles HTTP 429 (Too Many Requests) responses. When this status code is returned, the response will include a Retry-After header indicating how many seconds to wait before retrying. Your application should respect this header and delay its next request accordingly, rather than continuing to hammer the server. Failing to handle this gracefully may result in your requests being blocked.

If you have any questions or suggestions, don't hesitate to e-mail me or ping me on Matrix.

February 2026 News

· 6 min read
Brian Pepple
Founder of the Metron Project / Code Monkey

Monthly Statistics

During February the Metron Project added the following to its database:

  • Users: 152
  • Issues: 1,357
  • Creators: 246
  • Characters: 678

Thanks to everyone that contributed!

API Usage

Over the last 6 months, API usage has seen a significant increase — which is great! However, we are now at a point where our current server capacity is being strained.

Screenshot of a CPU load

Jonas Stendahl and I have landed some optimizations in the site's code, but they only go so far in relieving the pressure.

That leaves us with two options:

  1. Upgrade our server capacity — an additional $27–$43/month depending on the CPU tier selected.
  2. Reduce the daily rate-limit from 10,000 requests to something like 5,000.

Option #1 is clearly preferable, but we currently don't receive enough donations to cover the extra cost. That means we will likely need to reduce the daily rate-limit in the very near future, and we'll post an announcement here when that happens.

If you have other ideas or suggestions, please reach out!

Website Updates

Performance Improvements

Several changes were made to improve API and database query performance.

  • API query optimization: Refactored multiple API endpoint viewsets to reduce unnecessary database queries through better use of select_related and prefetch_related.
  • Additional database indexes: Added missing indexes on several models (Publisher, Series, Team, Universe) and their related issue querysets, which should speed up common lookups. (jyggen)
  • Fixed paginated Series queryset warning: Added explicit order_by() to SeriesViewSet and the publisher series-list action to resolve a UnorderedObjectListWarning that could surface during pagination.

API Enhancements

  • Conditional request support: The API now supports HTTP conditional requests using ETag and Last-Modified headers. Clients can send If-None-Match or If-Modified-Since headers to avoid re-downloading unchanged data, saving bandwidth. (jyggen)
  • Read dates in collection API: The collection list endpoint now includes read_dates data, giving API consumers access to reading history information directly from the list view.

Collection & Reading History

  • Reading history charts: The Collection Statistics page now includes daily (last 30 days) and monthly (last 12 months) column charts showing reading activity over time. Days with no reading activity are included in the daily chart.
  • Refined badge display logic: The read-count badge is now only shown when an issue has been read multiple times on the same calendar day, avoiding misleading counts from reads on different days.

User Interface Fixes

  • ESC key closes cover modal: On issue detail pages, pressing the Escape key now closes the cover image modal, matching standard UX conventions.
  • Variant formset display tweaks: Minor layout improvements to the variant issue formset to better display its fields.
  • User template improvements for editors: The user profile template now shows additional information for editors.
  • Series form validation fix: Handled a case where a missing series type would cause an unhandled error during form validation.

Other

  • Django 5.2.11 (#454): Updated to the latest Django 5.2 patch release.

Mokkari Updates

The following changes made to Mokkari, the Python wrapper for the Metron API, over the past month:

Added if_modified_since parameter to detail endpoints

All the detail methods (arc, character, creator, imprint, issue, publisher, series, team, universe, reading_list, and collection) now accept an optional if_modified_since datetime parameter. When provided, the request includes an If-Modified-Since HTTP header (formatted per RFC 7231) and returns None on a 304 Not Modified response — useful for efficiently detecting whether a resource has changed since you last fetched it. Naive datetimes are treated as UTC, and non-UTC datetimes are automatically converted before the header is sent.

Added Read Dates to CollectionList schema

The collection schemas have been updated to match the latest Metron API:

  • A new ReadDate model was added with id, read_date, and created_on fields.
  • read_dates and read_count fields were added to CollectionList and CollectionRead.
  • The date_read field type in CollectionRead was changed from date to datetime to align with the API's date-time format.

Migrated to PyrateLimiter 4.x

The rate-limiting implementation was updated for the breaking changes in pyrate-limiter 4.0:

  • Removed the now-deleted BucketFullException and LimiterDelayException imports.
  • Simplified Limiter construction.
  • Rewrote _check_rate_limit() to use the boolean return value of try_acquire(blocking=False) instead of catching exceptions.

Desaad Update

Here's a roundup of changes shipped to Desaad over the past month.

Reading History Import

The import_reading_status management command was upgraded: it now imports reading history dates from Metron, not just reading status. A new ReadDate model was added to store the actual date an issue was read, and the UserStatsView was updated to use these dates for reading activity stats. Previously, stats reflected the import timestamp rather than when issues were actually read.

Importer: Require External IDs

The comic importer was tightened up to require an external ID when creating characters, teams, universes, arcs, locations, genres, and creators. Previously, the importer would fall back to name-based lookups and create records without a verified external ID, leading to potential unverified duplicates. This change simplifies the importer significantly and ensures all entities can be reliably matched against a metadata source.

Reader UX: Double-Tap Confirmation

The comic reader's Prev Issue, Next Issue, and Close buttons now require a double-tap to activate. On the first tap, the button turns yellow and prompts "Tap again"; a second tap within 2 seconds proceeds with navigation. Otherwise, the button resets. This prevents accidental navigation away from your current reading position.

Infrastructure Updates

  • Python 3.14: The container image was updated to use Python 3.14.
  • Timezone via environment variable: The Django timezone setting is now configurable via an environment variable. Missing DB_* variables were also documented in the README.
  • Dependency updates: Project dependencies were refreshed across the board.

OpenCollective

A huge thank you to everyone who has contributed to our Open Collective! Your support makes a real difference in keeping the Metron Project running and growing.

What Your Contributions Support

Funds from Open Collective go directly toward:

  • Server hosting costs - Keeping the Metron website and API available
  • Domain registration - Annual domain name renewals
  • Future capacity increases - Scaling resources as the database and user base grows

All expenses are transparent and publicly viewable on our Open Collective page, so you can see exactly where every dollar goes.

Support the Project

If you'd like to help keep the lights on and support continued development, contributions of any size are appreciated and help ensure Metron remains a free resource for the comic book community.

Anyway, that's everything for this month! Take care.

January 2026 News

· 5 min read
Brian Pepple
Founder of the Metron Project / Code Monkey

Monthly Statistics

During January the Metron Project added the following to its database:

  • Users: 169
  • Issues: 2,238
  • Creators: 395
  • Characters: 1,430

Thanks to everyone that contributed!

Website Updates

New Features

Scrobble API Support

A new /scrobble endpoint allows users to quickly mark issues as read via the API with optional rating and timestamp. This enables integration with third-party reading apps and automated tracking workflows. The date_read field was migrated from a date to datetime for more precise tracking.

Reading History

Users can now view their reading history in a timeline layout with daily grouping. Recent reading activity also appears on user profile pages.

Multiple Read Dates

Collection items now support tracking multiple read dates, since comic issues are often re-read over time. The UI includes HTMX-powered controls for adding and removing read dates directly from the detail page.

Community View

User profiles now use username-based URLs (/accounts/username/) instead of numeric IDs. A new user list view with search functionality was added, making it easier to find and view other community members.

Improvements

HTMX Migration

Significant portions of the frontend JavaScript have been replaced with HTMX:

  • Global HTMX import in the base template
  • Navbar burger menu converted to HTMX
  • File upload handlers simplified with inline onchange attributes
  • Notification dismissals now use HTMX
  • Modal dialogs for read date deletion converted to HTMX
  • Removed several legacy JavaScript files (jquery.formset.js, bulma-modal.js, bulma-navbar-burger.js, bulma-notifications.js)

Auto-Dismissing Notifications

Non-error notifications now automatically disappear after 5 seconds. Error messages remain visible until manually dismissed.

Issue Autocomplete Search

The issue autocomplete now supports more flexible searching:

  • Search terms can be in any order ("Spider Amazing" matches "Amazing Spider-Man")
  • Series year filtering using parentheses (e.g., "Speed Racer (2025) #4")

Email Compatibility for Account Activation

Updated how Metron sends account activation emails to ensure they display correctly across all major email clients.

The changes include:

  • Multipart emails: Activation emails now send both HTML and plain text versions, ensuring users can read them regardless of their email client settings
  • Universal styling: Replaced CSS stylesheets with inline styles, since Gmail and Outlook strip CSS from the <head> section
  • Table-based layout: Switched from div-based layouts to tables for consistent rendering in older email clients
  • Better color support: Removed CSS gradients in favor of solid colors for Outlook compatibility
  • Web-safe fonts: Now using Arial and Helvetica instead of system fonts for cross-platform consistency

These updates should ensure new users have a smooth onboarding experience, regardless of whether they use Gmail, Outlook, Apple Mail, or any other email client.

Bug Fixes

  • Fixed duplicate issue creation race condition—users now see a friendly error instead of a 500 error
  • Fixed SlugRedirectView when requesting non-existent items
  • Fixed success message formatting in user account activation
  • Fixed sequential read date deletion (HTMX event handlers now reinitialize properly after content swaps)

Infrastructure

  • Updated to Django 5.2.10
  • Removed request for GitHub stars from the welcome flow

Desaad Updates

The following changes were made to Desaad.

Scrobbling to Metron

Issue scrobbling was added this month, and more information on it can be found in this blog post.

Database Migration to PostgreSQL

The application has migrated from SQLite to PostgreSQL. This followed an initial attempt to improve SQLite concurrency with WAL mode and retry logic, but PostgreSQL proved to be a better fit for handling concurrent operations during comic imports while the web server is running.

Reader Improvements

The comic reader received several updates:

  • Added previous/next issue navigation for reading through a series
  • Added gesture hints overlay to help users discover touch controls
  • Implemented browser caching for page images with proper HTTP cache headers (ETag, Cache-Control)
  • Improved performance by using the stored page count from the database rather than reading from the archive each time
  • Added customizable auto-hide delay for reader controls

After testing both approaches, the reader reverted to using base64 image data instead of image URLs, as this provided a better reading experience.

Background Task Processing

The app now uses Django 6.0's Tasks framework with django-tasks[rq] for several operations:

  • Reading list imports now run asynchronously with status polling via HTMX
  • Cover extraction is deferred to a background queue, making initial imports faster
  • Missing reading list items are automatically linked when matching issues are imported

Metron Integration

A new import_reading_status management command allows users to sync their reading history from their Metron collection. This imports completion status and ratings for issues that match by Metron ID. Currently, it only imports the read status, but in the near future it will add the dates the issues were read on also.

HTMX Adoption

Replaced JavaScript with HTMX in several places:

  • Notification dismiss buttons
  • Reading list detail page (delete modal, hide/show missing toggle)
  • Navbar burger menu

Other Changes

  • Added sorl-thumbnail with Redis cache backend for cover images
  • Added cover images to the series list view
  • Added a home view showing next unread issues for started series
  • Added completion indicators to reading list overview
  • Added django-debug-toolbar for development
  • Various container and deployment fixes

OpenCollective

A huge thank you to everyone who has contributed to our Open Collective! Your support makes a real difference in keeping the Metron Project running and growing.

What Your Contributions Support

Funds from Open Collective go directly toward:

  • Server hosting costs - Keeping the Metron website and API available
  • Domain registration - Annual domain name renewals
  • Future capacity increases - Scaling resources as the database and user base grows

All expenses are transparent and publicly viewable on our Open Collective page, so you can see exactly where every dollar goes.

Support the Project

If you'd like to help keep the lights on and support continued development, contributions of any size are appreciated and help ensure Metron remains a free resource for the comic book community.

Anyway, that's everything for this month! Take care.