Public MIT

Threads Automated Posting × SlimeTree-RLM mu-prefilter

Stop blow-up-bait posts before they go live

An in-browser RLM classifies your draft up-front — if mu, warn / suppress before posting; if D/R, post for real via the Threads Graph API. Access tokens stay in localStorage only and never traverse JAVATEL servers.

1. Threads access token

Create a Threads App at developers.facebook.com/docs/threads/get-started/ → "Generate Token" issues a long-lived token (60 days).
Required permissions: threads_basic + threads_content_publish.

Not authenticated (paste token and save, then click "Check user")

2. Post draft

Up to 500 characters. Before sending, SlimeTree-RLM classifies as D (deterministic fact) / µ (suppression candidate) / R (LLM judgment).

0 / 500 chars

RLM status: Not classified

Last verdict: —

2.5 LLM polish + blow-up risk check NEW: Pattern B (cross-vendor)

Have the LLM "score blow-up risk + improve phrasing" for your draft. API keys are shared with those saved in the Meta Gateway (localStorage). If none are set, go to the Gateway.

Single LLM — polish with one provider + one model Cost-tier (B) — cheap polish → judge → premium only for insufficient cases

provider: model:

cheap (B): 0 escalated (B): 0 esc rate (B): —

3. Post / API call log

(nothing has run yet)

Pricing example: 10,000 drafts processed/month

Monthly cost of the same workload across 6 LLM tiers. The RLM filters 73%, so the reduction ratio is the same on every provider. See the service page for details.

Tier	Class	No filter	With RLM	Savings/month	Annual
Gemini 2.5 Flash	Everyone	$7.50	$2.03	$5.47	$66
Gemini 2.5 Pro	Everyone	$103	$28	$75	$900
GPT-5 mini	Business	$26	$7	$19	$228
GPT-5	Business	$310	$84	$226	$2,712
Claude Haiku 4.5	Business	$42	$11	$31	$372
Claude Opus 4.7	Business	$1,530	$413	$1,117	$13,404

* 200 in / 1,000 out tokens. "No filter" = all 10,000 drafts go to the LLM; "With RLM" = only the 2,700 R-verdict drafts do. Prices reflect public unit rates as of 2026-05; industry mix is adjustable.

4. Why the RLM works for Threads auto-posting

µ verdict: Identity probes / future predictions / sensitive language are detected locally before posting → prevents lone-account suspensions caused by Meta's moderation
D verdict: Standard product announcements or fact-based posts are passed deterministically → raises the safety rate of automated posting
R verdict: Marginal cases are deferred to an LLM (Claude); the pre-chain record (WAL) still keeps them traceable
WAL audit: Every post candidate is recorded in a SHA-256 chain → you can later air-gap reproduce "who let that post through with what draft"

5. Source / License

All code in this demo (HTML + JS) is licensed under MIT, the same as the Gateway source page. The RLM mock uses the same slimetree-rlm-mock.js. The Threads Graph API call logic is in the inline <script> at the bottom of this page.

slimetree-rlm-mock.js (shared)
LICENSE.txt (MIT)
Threads API docs: developers.facebook.com/docs/threads/