What’s new in Aethrel.

Batch review with selection tools

Reviewers can now select multiple drafts at once using visible-set presets, preview each card before approving, and move through a batch with progress feedback. High-volume queues are dramatically faster to clear.

Keyboard shortcuts in the review queue

J and K navigate between drafts. A approves, R rejects. Reviewers who live in the queue can now move without reaching for the mouse.

Voice examples in onboarding

Clients can now provide concrete examples of language that sounds like their brand — and language they'd never use. These examples are applied during draft generation, not just as loose guidance.

Brand boundaries in onboarding

Never-engage topics, competitor names, and banned phrases are now captured during onboarding as structured inputs rather than free text. Hard limits are enforced at the pipeline level.

Weekly digest for Growth and Pro clients

A weekly summary of discovered conversations, replies posted, and engagement signals is now delivered automatically each week. Aethrels can also trigger a manual send from the settings panel.

Explain this draft

Reviewers can now request a plain-language explanation of why a draft was written a particular way — what thread signals it responded to, what constraints shaped it. Useful for catching edge cases and improving strategy guidance.

Source heatmap in client analytics

The analytics portal now shows a breakdown of activity by source — which subreddits and feeds are generating the most discovered threads, and which are producing the most approved replies.

Automation health dashboard

Aethrels now have a dedicated view showing whether each part of the pipeline — discovery, drafting, scoring, publishing — is running on schedule. Cron jobs emit heartbeats, and the dashboard surfaces gaps immediately.

Policy scoring surface in review queue

Policy scores are now displayed with the specific signals that contributed to them. Reviewers can see exactly what triggered a flag rather than just a score number, making edit decisions faster and more informed.

Dead-letter job alerts

If a pipeline job fails and can't be retried, operators now receive an email alert. Silent failures no longer go unnoticed.

Discover cadence timer fix

The discovery cadence timer was not resetting correctly after a manual queue clear operation. Fixed — discovery now resumes on the correct schedule.

Client analytics portal

Clients now have a dedicated portal showing discovered conversations, replies posted, engagement signals, and activity over time. The portal is intentionally scoped — clients see outcomes, not the internal queue or operator settings.

Onboarding wires up channels and strategy

Completing onboarding now automatically configures the primary posting channel and writes the strategy profile. Previously these required a separate setup step from the operator team.

In-app changelog

The operator console now surfaces a running changelog of platform updates so the team stays current without checking externally.

Review desk draft preview

The review queue now shows a character count on draft previews, making it easier to assess length before approving or editing.

Stripe billing — checkout, webhooks, and portal

Plan selection during signup now routes through a real Stripe checkout. Subscription state syncs automatically via webhooks, and clients can manage billing through the Stripe customer portal.

Multi-tenant workspace isolation

Every client workspace is now fully isolated at the data layer. Discovery, drafts, strategy, and analytics are scoped to the workspace and cannot bleed across accounts.

Workspace activation flow

Client signups now move through a clear pending → active state. Aethrels review each workspace before activation, and the status is visible in the client console and operator dashboard simultaneously.

Billing portal access from client settings

Clients can now access the Stripe billing portal directly from their settings page to manage plans, update payment methods, or view invoices.

LinkedIn adapter

Growth and Pro clients can now connect LinkedIn accounts and engage with LinkedIn feed posts. Tone guidance adapts automatically for LinkedIn's professional norms — drafts skew more direct and less casual than Reddit replies.

Account warm-up

New accounts now start with a conservative posting cadence — 1 to 2 replies per week — for the first 14 to 30 days. Pacing increases automatically as account history builds and platform standing is established. Existing accounts with prior activity have a shorter ramp.

Intent scoring v2

The thread relevance model was rebuilt from the ground up. Signal-to-noise on discovered threads improved substantially — less surfacing of technically-matching but contextually weak results.

Platform account health indicators

The operator console now shows health indicators for connected platform accounts — warm-up progress, recent posting activity, and any standing issues that need attention.

Human review queue

All drafts now route through a review step before publishing. Reviewers see the original thread, the draft reply, and its policy score. Approve, edit and approve, or reject — every action is logged.

Reddit OAuth connection

Aethrels can now connect Reddit accounts directly from the platform console using OAuth. No more manual credential entry.

Review assignment

Drafts in the queue can now be assigned to specific reviewers, preventing two operators from reviewing the same item simultaneously.

Thread deduplication

The same thread was occasionally surfacing more than once across back-to-back discovery runs. Fixed — each thread now appears at most once per active review cycle.

Context-aware AI draft generation

Drafts are now generated using the full thread context — the original post, reply chain, and community signals — alongside the client's strategy profile. Early drafts used the strategy profile alone; this significantly improved relevance and tone fit.

Policy scoring

Every draft now receives a policy score before entering the review queue. The score reflects how closely the draft stays within the client's defined boundaries. High-scoring drafts are flagged for closer reviewer attention.

Configurable discovery sources

Aethrels can now add, remove, and reprioritize discovery sources per client from the operator console. Previously, sources required a code-level configuration change.

Full audit trail

Every significant action — discovery run, draft generated, review decision, post published — is now logged with a timestamp, actor, and outcome. The audit trail is visible to operators and available for export.

Core pipeline: discover → draft → review → publish

The fundamental four-step pipeline is running end-to-end in a private environment. Discovery, draft generation, a basic review step, and publishing are all connected.

Reddit discovery in private testing

Subreddit monitoring is live for a small set of test sources. The discovery loop runs on a fixed schedule and surfaces threads matching configured topic signals.

Initial strategy profile

A basic strategy profile — brand name, description, and tone guidance — is captured at setup time and used to shape draft generation.