Waze Voicepack Links: an archive of voices Waze has quietly retired
A community-maintained archive of classic and custom Waze GPS voice navigation packs
At a glance
- What is it?
- This repository collects links to Waze voice packs, both the celebrity voices that were once officially contracted and the custom ones users recorded with the Voice Recorder feature, so they do not vanish with old Reddit threads. The important part is the October 2026 notice at the top: Waze changed how it serves these files and most of the links no longer work.
- Who is it for?
- Treat this repository as a directory to browse rather than a working source of voice packs right now, because the maintainer reports that Waze changed how these files are retrieved and most of the previously working links are no longer functional, with no known reason or workaround. It is still worth using as a map of which voices existed and where the official list of retired voices lived, and as the place to track what returns.
- Can I use it commercially?
- Yes, with conditions. GPL-3.0 is a copyleft licence: if you distribute software that includes it, you must release that software's source code under the same licence. Running it internally without distributing it does not trigger that obligation.
- Is it still maintained?
- Yes. The repository last received commits 4 days ago.
- What is it written in?
- Mainly Python, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on October 23, 2026, and from our analysis. They are not legal advice.
Editorial analysis
What the archive is for, and who keeps it
The repository is described as a community-maintained archive of classic and custom Waze GPS voices, all in one place, and the README puts the motivation plainly: bring back your favorite characters, celebrities and custom voice packs that have disappeared from the official app over the years.
The stated purpose of the collection is to gather publicly shareable Waze voicepack links so they do not get lost in old Reddit threads, dead websites or outdated blog posts. The three goals that follow are easy to find, easy to install and community maintained.
That last phrase carries more weight than it first appears, because Waze is a closed platform. There is no marketplace and no catalogue. A voice pack is a set of audio files that the app fetches from Waze's own servers and can install from a link, which means the only way to find one is to have a link, and the only way to keep one is to write it down somewhere. That is the gap this repository occupies, and it is a preservation task rather than a development one.
The licence is GPL-3.0, the last push was on 2026-10-20, and there are no GitHub releases, which is consistent with a repository whose content is a living table of links rather than a versioned artifact.
The primary language is recorded as Python, which comes from the tooling in the repository rather than the content. The requirements file lists pydub for audio handling and openai-whisper for speech recognition, alongside the usual requests and urllib3 stack.
The October 2026 notice, and why most links are now dead
The first thing in the README is an update, and it is the most consequential thing in the repository.
The notice says that within the last few days, many users reported issues with adding or downloading the majority of Waze voice packs uploaded using the Voice Recorder feature. It appears that Waze has made a change to the way these voice packs are retrieved from their servers, and as a result most of the previously working links below are no longer functional.
The README is careful about what it does not know. There is no known reason, no explanation and no workaround for this change at the moment.
That distinction matters, because it separates what the maintainer observed from what they concluded. The observation is that the links broke, that the break is recent, and that it affects the voice packs uploaded through the Voice Recorder feature in particular. The conclusion, marked as tentative, is that Waze changed the retrieval mechanism. The cause is unknown.
Two references are given for the wider discussion: a Reddit thread in the Waze community and issue 215 in this repository. If you are deciding whether to spend time on this project, those two links are where the current information is.
The practical consequence for a reader is that a table of links in this repository is not the same thing as a working set of voice packs, and the difference between the two is entirely outside the maintainer's control.
How installation works, and why it has to happen on the phone
The installation procedure is short and constrained by how Waze handles voice packs.
The steps are to install the Waze app on your phone, open the repository on the same device, tap a voice link from the list, and let Waze open and download the voice.
The second step is the one that trips people up. The repository has to be opened on the same device as the app, because the tap has to reach the Waze app as a link. Opening the list on a desktop and sending it to a phone by another route is not what the flow is built for, and the README does not describe a supported alternative.
The table of entries gives each voice two links plus notes. The first is an app-facing link on waze.com with an acvp parameter, and the second points at a tar.gz archive of the mp3 files on the voice-prompts host. A third column carries references where they exist, such as the Google blog post for the 80s aerobic instructor voice or the one for driving with Chaka Khan.
The mp3 column is worth understanding. Those are the raw audio archives, and they are the more durable part of the listing, since they point at the file host directly rather than going through the app-facing link flow. Whether they still resolve after Waze's change is not something the README answers, but it is where you would look if an app-facing link fails.
The archive also lists a couple of column values that are honest about quality. The official voice list section says the former contracted voices have been saved and converted into user-made custom voice packs with shareable links of varying quality.
The official voice list and the accent rows nobody expects
The largest section of the repository is the Waze Official Voice List, and it is not only celebrities.
The section covers current and former contracted celebrities and voice actors who at one point had their voices officially on the Waze app. A reading of the table shows a mix of licensed and in-house characters. The one-line entries include Boy Band, Cookie Monster, Dog and DJ 70's, and there are regional voice options such as British Brummie, Cockney and Mancunian accents, plus a Catriona Gray entry marked with a broadcasting role.
The named voices are the more recognisable half: Arnold Schwarzenegger for the Terminator, Batman voiced by Kevin Conroy, C-3PO from Star Wars, Colonel Sanders from KFC, Boy George, Christina Aguilera, Chaka Khan, Elvis Presley, Dizzee Rascal, Cupid, Ed Helms, and characters and performers tied to films and games including Andy Field from Five Nights at Freddy's, Armie Hammer as Jackson Storm from Cars 3, and an Alien.
The structure of the entries is uniform, which suggests the table is generated or maintained with some care rather than appended to casually. Each row carries a name, an app-facing link, a link to the mp3 archive, and a notes column that is often empty and occasionally a reference.
The Wikipedia and Waze accounts of why these voices disappeared are not in the repository, and the README does not attempt to explain each removal. The stated framing is simpler: they were official once, they are gone now, and the community has converted them into user-made packs so the voices can be kept.
The tooling around the link list, and how a pack is put together
The repository is a table of links, but it is not only a table of links.
The top level holds an assets directory, a contributing.md file, a helper_files directory, an mp3_upload directory and the requirements.txt. The mp3_upload directory is the interesting one, because it implies the workflow goes further than collecting links: somebody produces the audio and uploads it.
The dependency list gives the shape of that workflow. pydub is an audio manipulation library, and openai-whisper is a speech recognition model. Together they suggest a transcription step, which makes sense for a voice pack: the prompts have to match what the app expects to display, so the recorded audio needs to be transcribed and checked against the expected vocabulary. protobuf and blackboxprotobuf come along with the Waze voice pack format, and requests and urllib3 handle the upload and download.
The contributing.md file is where the workflow is presumably described, and it is one of the files a new contributor should read first.
The requirements are pinned exactly, with pydub at 0.25.1, requests at 2.32.3, urllib3 at 2.2.3, protobuf at 3.10.0 and setuptools at 75.2.0, and openai-whisper listed without a version pin. That is the sort of pinning you want for a tool that interacts with an external service whose format is not documented publicly, since an unpinned transitive change can break the upload flow without anything in the repository changing.
For most readers none of this matters, because the README is the product. The tooling exists to serve the community that keeps adding packs to the list, and to keep the list honest as Waze's behaviour changes underneath it.
Editorial conclusion
Treat this repository as a directory to browse rather than a working source of voice packs right now, because the maintainer reports that Waze changed how these files are retrieved and most of the previously working links are no longer functional, with no known reason or workaround. It is still worth using as a map of which voices existed and where the official list of retired voices lived, and as the place to track what returns. Before you spend time trying to install anything, read the UPDATE section and issue 215, and check the accompanying mp3 files links, which point at the voice-prompts host directly and may behave differently from the app-facing links.
Frequently asked questions
Do the Waze voicepack links in this repository still work?
The October 2026 update says that most of the previously working links are no longer functional, because Waze appears to have changed how these voice packs are retrieved from its servers. The maintainer states there is no known reason, explanation or workaround for the change at the moment.
How do I install a Waze voice pack from the list?
Install the Waze app on your phone, open the repository on that same device, tap a voice link from the list, and Waze opens and downloads the voice. Opening the list on a desktop is not the flow the procedure is built around.
What is the difference between the two links in each voicepack row?
One is an app-facing link on waze.com with an acvp parameter that triggers the install in the app, and the other points at a tar.gz archive of the mp3 files on the voice-prompts host. Some rows also carry a reference such as a Google blog post about that voice.
Where is the discussion about the broken voicepack links?
The README points to a Reddit thread in the Waze community and to issue 215 in the repository itself. Those are the two places where the current information about the change is being tracked.
What does the official voice list in the repository cover?
Current and former contracted celebrities and voice actors who at one point had their voices officially in the Waze app, from Arnold Schwarzenegger and Kevin Conroy as Batman to Boy George, Chaka Khan and Colonel Sanders, alongside regional accent options. The README notes the voices have since been converted into user-made custom packs of varying quality.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/pipeeeeees-waze-voicepack-links)Was this useful?