CalmSource
§ 01 — Overview
Folio № 01 · OverviewVancouver · MMXXVI

The Unified Edge-Aggregation Engine

for connected media platforms.

§ 1.2 — Thesis

Problem · Solution · Why Now · Ask

ProblemP
Aggregation on constrained hardware is broken.
The installed base of connected media devices is dominated by 1 GB Android sticks, budget Smart TVs, and legacy set-top boxes. Every OEM and operator rebuilds the same aggregation glue — badly — and users pay for it with OOM crashes, buffering, and stale catalogs.
SolutionS
A memory-aware, on-device edge runtime.
CalmSource replaces that glue with a single Android runtime: decentralized manifest resolution, throttled feed indexing, and a stream-racing QoS router — all capped against measured device RAM so OOM is eliminated by construction.
Why NowN
Local-first is becoming the default.
Regulators and platform owners are converging on on-device processing; server-side session models are getting harder to defend. A runtime that never creates one is a durable position, not a temporary workaround.
The AskA
Pre-seed SAFE. First check.
Post-money SAFE, size scoped with the lead. Directional use of funds: 12 months of founder runway plus the first Android runtime engineer hire. Not a fixed ask — right-sized in the first serious conversation.
01§ 2.1

Decentralized Extension Hub

An open, asynchronous JSON manifest pipeline. Resolves community-driven metadata providers and distributed catalogs concurrently at the edge — no centralized database overhead.

02§ 2.2

Asynchronous Feed & EPG Synchronizer

A localized high-throughput write path. Throttled batch transactions and memory-aware background workers index tens of thousands of streams without pressuring the device.

03§ 2.3

Algorithmic Quality-of-Service Router

A stream-racing network engine. Parallel headless probes measure latency and hardware codec parameters, then commit to the optimal data path in milliseconds.

Architecture Focus

Memory-Aware Edge Buffering

Real-time data buffers are capped against measured device RAM. Out-of-Memory kernel panics on constrained media sticks and legacy hardware are eliminated by construction, not by hope.

Fig. 01 — Buffer ceiling / device classn = 4 tiers
Legacy Stick512 MB38% / cap 42%
Entry TV1 GB55% / cap 62%
Mid Phone3 GB71% / cap 82%
Flagship8 GB+46% / cap 94%

Compliance Focus

Privacy-by-Design Framework

Credential mapping, indexing, and token management stay inside the device's secure storage. Log lines are redacted before they can leave the process. No server-side session is ever created.

Fig. 02 — Local isolation pipelinescope: on-device
  1. 01
    Credential Vault
    AES-GCM · Android Keystore
    local
  2. 02
    Index & Token Map
    Encrypted at rest — never synced
    local
  3. 03
    Log Redactor
    Automated PII scrubbing
    local
  4. 04
    Outbound Telemetry
    Disabled by default
    local

// no server-side session is ever created

Pre-Seed · Looking for the first check

One founder, a working prototype, and a small closed beta.

No prior investors, no revenue yet. What exists today: a working Android runtime on real low-RAM hardware, a small group of beta testers, and a specific plan to turn a technical wedge into a product.

Stage
Pre-seed
Prior investors
None
Team size
1 founder
Product
Closed beta
Instrument
SAFE (post-money)
Read the Pre-Seed BriefRequest an Intro Call