T-HUB/docs/decisions/2026-10-09-intake-posture.md
glass beadian aa5e262f3a Tactical Media Broadcast Hub — initial release
Self-hosted multi-platform broadcast newsroom for activist collectives:
publishing pipeline with group approval and DST-proof scheduling, TweetDeck-style
column wall with hostile-content-safe rendering, live control room with Auto
Director, bridged chat governance with polls and quorum, distributed sandboxed
transcode workers, encrypted content-addressed archive, monthly analytics.

MIT licensed. See README.md and docs/.
2026-10-10 11:56:42 +00:00

22 lines
2.6 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

# Decision: Content intake posture — 2026-10-09
**Decision:** metadata-only + community submissions + Tier-1 sanctioned sources. **No gray-market account farming.**
## Context
The collective wants an intake firehose: continuous content ingestion from social platforms for media work — hashtags, accounts, search — feeding Streams columns, Compose, and the archive.
## Research base (live-verified 2026-10-09; the research archive lives outside this repository)
- `2026-10-09-intake-firehose-architecture.md` — per-source feasibility, verified volumes (Jetstream ~52 posts/s measured, Mastodon 300 req/5min anon verified, YouTube channel RSS keyless verified), architecture, storage math (~37 GB/day modeled), phased rollout.
- `2026-10-09-scraping-feasibility.md` — farm-account + browser-automation scraping: every platform rated Brittle or worse; enforcement record (X C&D'd Nitter Aug 2026; Meta won injunctions specifically on fake-account scraping).
- `2026-10-09-gray-market-services.md` — the vendor layer exists and is priced ($1–50 aged accounts, SMS pools, captcha farms, proxies, anti-detect browsers, pre-scraped datasets) but delivers +15–20% coverage for 10–30× the cost of the alternative posture, with 60–90% farm-detection probability within a year, material contagion risk to legitimate accounts, and documented identity-crime supply chains.
## Decision
1. **Tier 1 (build now, zero ban risk, no accounts):** RSS/Atom (incl. YouTube channel RSS), Mastodon public tag/account timelines, Bluesky Jetstream with server-side keyword gate, PeerTube REST, per-URL public fetchers (gallery-dl/instaloader-class).
2. **Tier 2 (small bounded pool, later):** 1–3 Telegram MTProto session accounts for public channel reading (quiet, warmed, never org-linked); optionally 3–5 read-only X accounts, dedicated egress each, 90-day sunset, kill switches.
3. **Rejected:** mass account farming, hashtag firehoses via farms, gray-market datasets, managed-scraping retainers — 10–30× cost, existential tail risk, documented identity-crime supply chain (see gray-market ledger §9.4).
4. **Per-URL ingestion** covers locked platforms: community-submitted public post URLs are fetched individually (gallery-dl/instaloader-class) — categorically different from authenticated crawling.
## Consequences
- Intake adapters draw from `packages/egress` per-host budgets.
- Locked-platform columns show community-submitted + metadata-only content; full firehose coverage concentrates on sanctioned sources.
- Decision revisitable if platform APIs open or a funded data vendor passes legal review.