ReplayLab Upload a demo
  1. ReplayLab
  2. Features

CS2 voice chat from a demo, synced to the replay

Last updated: November 11, 2026

A lot of what decides a round is said, not shown: the call for a rotate, the “one B, he's low”, the silence before a bad push. If your CS2 demo recorded in-game voice, ReplayLab pulls it out of the .dem file, splits it by team and by round, and plays it in sync with a 2D replay of the match — in the browser, without launching the game.

What voice data in a CS2 demo actually is

When a server records voice into a demo, it stores the in-game voice chat as a stream of small compressed audio packets. Each packet is tied to the player who was speaking and to the game tick when it was sent. That is enough to rebuild who said what, and exactly when, relative to everything else that happened in the round.

Only the game's own voice chat ends up in the demo. Calls made over Discord, TeamSpeak or any other app are never recorded there, so no tool can extract them from a .dem file.

Which demos contain voice

Not every CS2 demo has voice. Whether it is there depends on the server that recorded the demo and how recording was configured. Some platforms and servers keep team voice in their demos; many demos, including plenty of matchmaking demos, have none at all. A demo that was recorded without voice cannot get it back later — the audio simply is not in the file.

How to check: upload the demo. If it contains voice, the replay shows a Voice button; if the button is not there, the demo has no voice packets that could be matched to the two teams.

How ReplayLab extracts voice from a demo

Voice extraction runs as part of the normal demo analysis, so you don't need a separate tool or upload:

  1. Read the voice packets. The parser reads every voice packet together with its tick and the SteamID of the speaker.
  2. Assign speakers to teams. Each speaker is matched to one of the two teams of the match. Packets from people who were not on either team (for example, spectators) are left out.
  3. Cut by round. For every round, the audio starts at the beginning of the round — the same zero point as the replay timeline — and runs until the official end of the round plus four seconds, so the post-round talk after the last kill is kept.
  4. Mix one track per team. All players of a team are decoded and mixed into a single mono track for that round, then encoded as Opus audio. A track is only created if someone on that team actually talked during the round.

The result is up to two short audio tracks per round — one for each team — that line up with the replay to the second.

Listening to comms in the 2D replay

In the replay, the Voice button switches between four modes: off, team A only, team B only, and both teams together. The same choice is available in the replay settings. Your choice is remembered in the browser for the next match.

  • Synced to the timeline. Seeking in the replay jumps the audio to the same moment. While voice is playing, the replay follows the audio clock, so what you hear matches what you see on the map.
  • Round by round. When you switch rounds, the matching track is loaded; the next round's audio is fetched in advance.
  • Normal speeds only. Above 2× playback speed the voice is muted, because speech becomes unintelligible.
  • Works on phones. The site is built mobile-first. On iPhone, playback is set up so that the hardware silent switch does not mute the comms.

What to listen for when you review comms

Voice is most useful next to the moments where a round was lost. A practical routine:

  • Open a death flagged in the positioning breakdown, jump to that moment in the replay and listen to the ten seconds before it. Did anyone call the enemy you died to? Did you call your own position?
  • Compare the calls with the map. If someone says “three A” and the 2D replay shows two, the info was wrong or late — that is worth knowing.
  • Listen to the other team when their voice is recorded too. Hearing how an opponent's execute was called often explains why your setup broke.
  • For team reviews, watch together: ReplayLab has a shared watch room where the whole team follows the same replay, and everyone can turn the comms on.

Privacy

Voice is personal data, so it is treated like the rest of your match. Only you can see your uploaded matches — and your team, if you join one. The audio is served only to people who can open that match. You can delete a match, and with it the extracted voice, at any time.

Limitations

  • No voice in the demo means no voice in the replay. ReplayLab cannot recover audio that was never recorded.
  • Only in-game voice chat. Discord and other external voice apps are not part of a demo.
  • Each team is one mixed track. You can switch between teams, but not mute a single player.
  • There is no separate audio download button at the moment — the comms are played inside the replay.

FAQ

Does every CS2 demo contain voice chat?

No. Voice is only in the demo if the server recorded it. Many demos, including a lot of matchmaking demos, have no voice at all. If your demo has voice, the ReplayLab replay shows a Voice button.

Can I hear Discord calls from a demo?

No. A CS2 demo only contains the game's own voice chat. Discord, TeamSpeak and other external apps are never recorded into the demo file.

Can I listen to only one team?

Yes. The Voice button switches between off, team A, team B and both teams. Each team is a single mixed track per round, so you cannot isolate one player.

Is the voice synced with the replay?

Yes. Each round's track starts at the beginning of the round, the same zero point as the replay timeline. Seeking moves the audio, and while voice plays the replay follows it.

Can I download the audio as a file?

Not at the moment. ReplayLab plays the comms inside the replay; there is no separate export button.

Is voice included in the Free plan?

Voice is part of the replay of every analysed match. The Free plan ($0) lets you explore sample analyses and analyse your first own demo free; further demos require Pro.

Read next