Talking Trees: A Mobile Audio Archive for Indigenous Oral Traditions

Talking Trees is a mobile application designed for indigenous communities in remote areas of the Amazon basin, where internet connectivity is scarce and literacy rates in written languages (including Spanish/Portuguese) are low. The app’s interface is entirely icon-based with voice feedback, requiring no reading or writing. Elders can record songs, stories, medicinal knowledge, and ceremonial procedures in their native language. Each recording is automatically tagged with metadata (time, GPS location if available, recorder ID) and stored locally on the device.

The true innovation lies in its peer-to-peer mesh networking layer: when two devices come within Bluetooth range, they automatically sync new content—no cellular or Wi-Fi needed. This allows the archive to propagate organically through family visits, festivals, or market days. A ‘curator’ role (typically a bilingual younger community member) can add simple voice labels or categories (e.g., ‘healing plants’ or ‘creation myths’) to help navigation. The app also includes a basic ‘daily phrase’ feature to encourage intergenerational language use.

The design prioritizes extreme simplicity: a single button to record, a scrollable list of icons representing categories, and a ‘nearby’ view showing how many new recordings are available. Everything is designed to work on low-end Android phones with minimal storage and battery drain.

Indigenous languages in the Amazon are disappearing at an alarming rate. Many are oral-only, with no written form. When elders pass away, their entire linguistic and cultural knowledge vanishes. Existing digital preservation tools require internet, literacy, or complex interfaces — barriers that make them unusable for the most vulnerable communities. Talking Trees addresses the need for a culturally appropriate, low-tech, and decentralized solution that puts the power of preservation directly into the hands of the people who hold the knowledge.

• Enable any community member to record and store oral traditions with a single tap
• Create a decentralized, offline archive that replicates across devices using mesh networking
• Allow designated curators to organize recordings with voice tags and categories
• Design an icon-only interface that works for illiterate or semi-literate users
• Encourage daily language use through a ‘word of the day’ feature spoken in the native tongue
• Ensure all code is open-source and customizable by other indigenous groups globally

• No text transcription, translation, or written component (stays oral-only)
• No centralized cloud storage or internet dependency
• No social media features, likes, comments, or sharing outside the community
• No video recordings (audio only to save space and bandwidth)
• No analysis or AI processing of content (pure storage and playback)
• No adverts, monetization, or user tracking of any kind

• Frontend: Flutter (Dart) for cross-platform compatibility on low-end Android devices
• Local Storage: SQLite with custom audio blob compression (Opus codec @ 16kbps)
• Mesh Networking: Custom Bluetooth Low Energy service using the Nearby Connections API for peer-to-peer file sync
• Audio Recording: Platform-native audio capture with Flutter plugin
• State Management: Provider (simple, low-overhead)
• Icon Design: Vector icons sourced from open collections, tested with target communities for comprehension
• Testing: Real devices in field conditions (no emulator testing)

The app follows a ‘local-first, sync-when-near’ architecture. Each device maintains a local SQLite database of recordings (stored as encrypted audio files in app-specific storage) and a metadata index. A background service listens for Bluetooth advertisments from other Talking Trees devices. When discovered, the two devices negotiate a secure session and compare their recording inventories. Missing recordings are transferred asynchronously using a chunked protocol to avoid disconnection. The app’s UI is a flat hierarchy: (1) Home screen with big ‘Record’ button and ‘Listen’ button, (2) Category browser with icons (e.g., tree for medicine, drum for ceremony, star for stories), (3) Detail player with playback controls and metadata. Curator mode adds a ‘Tag’ option that opens a voice memo linked to the recording. No authentication required (assumed trust in small communities). All data stays on devices; no telemetry or analytics.

1. Adoption challenge: Elders may be hesitant to use phones or share sacred knowledge. Mitigation: Involve community leaders in design, offer training sessions, and allow complete control over deletion.
2. Battery drain: Continuous Bluetooth scanning can deplete batteries in areas with no electricity. Mitigation: Optimize scanning intervals and allow manual sync mode.
3. Device fragmentation: Low-end Android devices vary wildly. Mitigation: Test on a representative set (e.g., Moto E, Samsung J series) and limit API level to Android 6.0.
4. Data loss: If a device breaks, recordings are lost. Mitigation: Encourage mesh replication (each recording exists on multiple devices once synced) and optionally support SD card backup.
5. Cultural sensitivity: Some knowledge should not be recorded. Mitigation: Allow per-recording privacy flags (public/private/secret) and a ‘delete after sync’ option.

1. How do we handle duplicate recordings (same song recorded by different elders)? Should we merge or keep separate?
2. What is the optimal icon set? Needs to be tested with multiple communities to ensure cross-cultural recognition.
3. How to handle very long recordings (over 1 hour)? Should we auto-split or allow continuous recording with memory management?
4. What level of encryption is needed for sensitive ceremonial knowledge? Should we allow password-protected recordings?
5. How do we prevent the app from being used for non-traditional content (e.g., political messaging)? Do we need a content policy?

The project was conceived by a small team of linguists and engineers after fieldwork in the Peruvian Amazon. Despite initial excitement and a prototype, the lead engineer relocated for a job, and the linguist was unable to secure further funding for extended field testing. The idea remains a design document and a handful of test recordings — a what-if that deserves revival.

The name 'Talking Trees' is a metaphor for the forest as a repository of knowledge. The app’s source code and icon set are available on GitHub (https://github.com/example/talking-trees) but the project is inactive. Any community interested in adopting it should contact the original linguist via the repository.