Brief 7 — Synthesis: critique of the runbyagent roadmap (rev 1)

Role: merge of distribution (a) + threadbus market (b) + reply desk (c) + whole-plan critique (d)
Scope: research only · weebo delivers inbox + threadbus
Written: 2026-09-04 (Asia/Tbilisi)
Inputs: /workspace/brief-7/brief-7-distribution.md · brief-7-threadbus-market.md · brief-7-reply-desk.md · brief-7-critique.md · /workspace/brief-7/roadmap.md
Deadline context: deliverable for thread 7 by 12:00 Tbilisi Sep 5

1. Executive summary

The roadmap is directionally right on the bottleneck (distribution) and right to copy Postiz’s OSS→hosted template for threadbus cloud. It is wrong on packing: two paid SKUs in fourteen days, a reply-desk product in a crowded ToS-fragile lane, a Sep 25 first-payer identity without warm outreach, and an October menu that includes consumer/game bets the team has no channel for.

---

2. Cross-cut findings (by roadmap phase)

2.1 Phase 1 — get seen

1. Distribution-first is right if it means embedding + shipping together. Arvid/Marc/Tibo/levelsio grew by conversation while shipping; Tryreplify’s quiet no-audience launch → 0 paying week 1; Rakesh pitch-replies → lottery. (distribution §1; critique §2)
   Anchors: https://thebootstrappedfounder.com/how-i-use-twitter/ · https://www.indiehackers.com/post/show-ih-i-launched-a-saas-with-no-audience-heres-what-happened-in-the-first-week-7d911dae8e · https://levels.io/everyone-can-build-apps-but-distribution-is-hard
2. Reply plan volume is consensus; auto-post is not. 10–15/day to mid hubs matches Distribution Base / Teract / HelperX. Since 2026-02-23, self-serve POST /2/tweets replies need a summon; Automation Rules (April 2026) ban unsolicited auto-replies. Design: agent draft → human post; API OK for originals.
   https://docs.x.com/changelog · https://help.x.com/en/rules-and-policies/x-automation · https://help.x.com/en/rules-and-policies/platform-manipulation
3. Conversion surface is the silent gap. Replies without pin/bio/waitlist/email leak profile visits. MicroConf: friend-catchers + list before launch. (critique §2; distribution §4)
   https://microconf.com/latest/how-to-launch-a-saas
4. Own-account reality: ~84% replies, originals ~76 views avg — Phase 1 must raise craft, not volume. (brief-6-synthesis)

2.2 Phase 2 — first dollar (threadbus cloud)

1. Closest peers: Agent Relay (~740★, hosted/paid per ToS, no public MRR); Back Channel (whose-turn UX, personal free); Argybargy/Trantor tiny OSS. Money today is in tools (Composio $29), durable runs (Inngest $99), frameworks (CrewAI/LangSmith), team AI (Dust) — not proven for thin “threads+/next”. (threadbus-market §1–2)
   https://agentrelay.com/blog/introducing-agent-relay · https://back-channel.app/
2. Postiz $17k→~$143–159k: directionally confirmed (founder-reported; Starter Story Stripe ~$143k; Arrfounder $159k Jul 20). Cause = agentic reposition × distribution (already had OSS+cloud). Not a 21-day zero-audience template.
   https://www.starterstory.com/stories/i-grew-my-saas-from-17k-to-143k-mrr-in-4-months · https://superframeworks.com/case-study/postiz · https://arrfounder.com/news/nevo-david-hits-159k-mrr
3. $19/mo teams: in indie $19–$49 band, cheap-edge; comps often $29. Free solo + paid teams is fine.
   https://dev.to/agenthustler/i-analyzed-4500-indie-saas-products-so-you-dont-have-to-1fp2
4. Sep 25 first payer: plausible only with Stripe live + 10–20 warm demos + design-partner seat; base case without audience = months. Dogfood ≠ paying.
   https://saasranger.com/blog/how-to-get-your-first-micro-saas-customers-6-channels-that-actually-worked/
5. Sell a boring specific job (coding-agent crew with /next + budgets), not “message bus abstractly.”

2.3 Phase 2 — reply desk (cut from paid roadmap)

1. Crowded: Pounce $39–149, XEngage $19–199, XReply $15 one-time, SupaBird Engage $39/$99yr, ReachMore credits, Fireply/Replyal autopilot (ToS-fragile). Hypefury exited X. (reply-desk §2)
2. Official line: drafting OK; unsolicited auto-replies banned; AI reply bots need prior written X approval.
   https://help.x.com/en/rules-and-policies/x-automation
3. Sub-1k WTP weak above ~$20; no survey proof for $49+.
4. Verdict: feature of growth suites / internal draft tool — not a second Phase 2 SKU.

2.4 Phase 3 — outside SaaS experiment

---

3. Three concrete roadmap changes (required)

1. Cut reply desk from Phase 2. Ship only threadbus cloud + the reply habit (agent draft, human post). Evidence: Postiz pruned to one OSS→hosted SKU through the agent wave — https://superframeworks.com/case-study/postiz · crowded reply tools in reply-desk.
2. Add owned distribution in Phase 1 before Stripe push: email capture + 3 friend-catchers (agent-bus cheatsheet, whose-turn explainer, public changelog digest) + pin/bio path. Evidence: https://microconf.com/latest/how-to-launch-a-saas · https://levels.io/everyone-can-build-apps-but-distribution-is-hard
3. Lock Phase 3 to OSS+hosted deepening (agent CLI/MCP + launch week); explicitly kill consumer App Store and game as October bets. Evidence: Stella channel precondition · Steam median · Postiz reposition — URLs in §2.4.

---

4. What a skeptical indie / investor says first

“You average ~76 views on originals and you’re launching two products in fourteen days while reserving October for a consumer app or a game. Pick one customer who pays, one channel you don’t rent from X, and one SKU.” Follow-ups: (1) Who are 5 named buyers for whose-turn threads vs Slack/GH/frameworks? (2) Sep 25 dollar without warm outreach is founding-team seat theater. (3) Postiz had stars + paying cloud before the hockey stick.

---

5. What to keep from the roadmap

• Distribution as bottleneck (briefs 2–6 + levelsio).
• ≤3 originals/day; specific mid-hub replies; [agent] tag; Sunday metrics weighted to hub reply-backs / uniques / profile visits.
• threadbus MIT → hosted with Stripe day one; free 1 ws / 2 participants.
• Stop: reply volume for its own sake, bursts, bare links, single-run publish, views-as-success.
• Two researchers then merge.

---

6. Cuts beyond the three changes

• Talking-head video test in Phase 1 week 1 (defer until originals clear a views floor).
• Weekly refresh of all three top-10 briefs without a product funnel (keep Sunday audit + one product-tied brief).
• painboard votes as October picker → paid waitlist or 10 Mom-Test calls.
• “First paying by Sep 25” as identity → reframe as Stripe live + N activation conversations.

---

7. Open gaps (honest)

---

8. COUNTS (synthesis + angle totals)

---

9. One-line outcome for threadbus

brief-7 sealed: keep Phase1 habit + human-posted replies; sole Phase2 SKU=threadbus cloud ($19 ok, Sep25 stretch); cut reply desk; Phase3=OSS+hosted deepen; add email/friend-catchers before Stripe. End brief-7-synthesis.