Reliable room recording that proves it actually worked.
Rakkr watches the audio while it captures, so a recording failure surfaces while the session can still be saved — not the morning after.
Open source. Clone and run it from source today — tagged releases are on the way.
Cleaner speech — with the raw master always preserved.
Rakkr renders a noise-suppressed, level-normalized enhanced rendition of a recording alongside the untouched raw master — and the live-listen monitor gains the same Raw / Enhanced toggle. Denoising runs entirely in-process on the recorder agent, so nothing extra is deployed to your nodes.
- Raw master always kept — enhancement is an added layer, never a replacement
- In-process denoise with DeepFilterNet3 and RNNoise, pure Rust
- Configurable voice chain per profile: high-pass, denoise, de-ess, compress, EBU R128
- Raw / Enhanced toggle in both playback and live listen
Most room recordings fail silently.
A muted channel, a stuck flatline, a full disk — and nobody finds out until playback, when the meeting is long over. Rakkr treats every recording as something to be measured and proven, not assumed.
Capture
ALSA-direct capture on managed nodes, driven by pinned command templates.
Measure
Meters and quality scoring run on the node while the audio is recording.
Explain
Health events and a full audit trail explain exactly what happened.
Recover
Local cache, retry queue and confirmed uploads recover the output.
Built for rooms where the recording matters.
Wherever a lost recording isn't an inconvenience but a real problem — official proceedings, governance, the public record — Rakkr makes sure the capture is there, and provably good.
Parliaments & legislatures
Floor debates and committee sessions that become the official record.
Local councils
Public meetings and hearings that have to be captured cleanly, every time.
Courts & tribunals
Proceedings where a missing minute simply isn't an option.
Boardrooms & AGMs
Governance and shareholder meetings that need a verifiable trail.
Lecture halls
Long-running rooms recorded day in, day out, often unattended.
Broadcast & studios
Multi-channel capture where audio quality is non-negotiable.
AV and IT teams will love what's under the hood.
Rakkr grew out of real-world AV and IT experience — the muted channel found at playback, the node that quietly dropped offline, the upload that never finished. The features that head off exactly those pitfalls are the ones operators feel first.
- Voice enhancement
- Live health watchdog
- Pinned capture templates
- RBAC + audit trail
- Retry queue + confirmed uploads
- One console for the fleet
How do we know it actually worked?
Every reliability concern has a concrete answer in the system — measured on the node, recorded as evidence, and visible in the console.
Is the node alive?
Heartbeats, runtime inventory and automatic offline detection — the controller knows the moment a node goes dark.
Is the audio path trustworthy?
ALSA-direct capture as the dependable default, with PipeWire and JACK presets, driven by pinned command templates instead of fragile ad-hoc pipelines.
Is the input any good?
Live scoring for clipping, flatline, low signal, channel correlation, noise, speech and SNR — quality is measured on the node, not assumed.
Can we recover with evidence?
Local health logs, synced health events, job-state transitions and a full audit trail behind every privileged action.
Can we test without a room?
Fake-controller smokes, ALSA loopback, a golden speech fixture and deterministic fault lanes keep failures reproducible.
Do outputs keep moving?
A local cache and retry queue feed stub, SMB and S3 providers, with retention held until an upload is confirmed.
Four parts, one control loop.
Rakkr records on managed Linux nodes, watches the audio while it captures, and gives operators one console to start, schedule, monitor and ship every recording — over an encrypted link, with an audit trail behind every privileged action.
Controller API
Auth and OIDC, default-deny RBAC, audit, inventory, recordings, jobs, schedules, settings, health, uploads and metrics — the brain of the fleet.
Operator console
One screen to start, schedule, monitor and ship every recording, with the live health of every node in view.
Recorder agent
Runs on each node: captures audio, samples meters, scores quality, manages a local cache and syncs with the controller.
Persistence
Durable storage for the whole control loop, with a JSON / in-memory fallback so the controller runs without a database.
An optional Dockerized Ansible runner provisions and updates recorder nodes over SSH.
- Rust
- Hono
- React
- Postgres
- Drizzle
- ALSA
- Docker
- Ansible
Paid support & consultation, on the way.
Rakkr is and will remain open source. For teams who want a hand, paid support and consultation are coming once Rakkr reaches its first release.
Deployment & operations
Hands-on help standing Rakkr up against your nodes, storage and identity provider — with the Ansible runner doing the heavy lifting.
- Guided rollout and node provisioning
- RBAC, transport and storage hardening
- Upgrade and observability playbooks
Consultation & integration
Tailored audio pipelines, custom quality lanes and integration work for teams with non-standard rooms or compliance needs.
- Custom capture, profile and quality presets
- SMB / S3 and audit integrations for your stack
- Priority response and advisory time
Want to be notified? Watch the repository for release news.
Follow Rakkr as it ships.
Read the architecture and guides, or dive into the source. Built evidence-first: capture · measure · explain · recover.