SM Autoposter

SM Autoposter

Ready

Secure Access

Sign in to continue

Use your team account before opening client data, drafts, sources, or reports.

Internal Deploy Agentic workspace
Client data stays hidden until sign-in completes
Publishing remains human-approved

Stage 1 of 5

Choose the source.

Attach the material this post should use. Everything else stays out of the way.

Source (attached)

What are we working from?

Add or change source

Source input Use this source for Trusted source

Post link

Destinations

More options

Purpose

Style

Brief What should this post do?

Master post (editable)

0 / 1,300

Exact outgoing proof

This is exactly what will be sent.

Source-link compliance

Compliant Links must come from the attached source or approved client destination.

Human approval requiredNothing will be published until you approve the final post.

Review lock

Final approval is required before anything is sent or scheduled.

Draft settings

Post type Source input Post link Add source link to post Link target Choose where the post should send readers. Trusted source

Send timing

When should it go out?

Finish approval to unlock sending.

Timing Date and time

Command plan

No command plan yet.

Clients 0

Active client profiles

Drafts 0

Posts in workflow

Scheduled 0

Publishing queue posts

Lessons 0

Stored improvements

System Status

Configuration

Auth Diagnostics

Check Clerk session health without exposing tokens.

Beta Operations

Beta Readiness

Hosted draft smoke command Copies a one-time command from your current Clerk session. Do not paste the token into chat.

Sign in first, then copy the command.

Beta Operations

Test Data Cleanup

Client Workflow

Next Step Path

Operator Checklist

Client Readiness

Client Intake

Client profile workspace

Generated profile you can edit

Voice & Guardrails

Selected client Business Website URL Audience Offers Forbidden claims Voice examples Approval rules

Humanization profile

Use these as editable style controls for draft generation and posting output validation.

Reading grade target Jargon to use Fluff to avoid Sensory / personal details CTA frequency Soft CTA templates

Prompt settings

Proposal-applied guidance for future draft generation. Operators can adjust it manually after reviewing performance evidence.

Recommended posting style Prompt notes

Add more resources, blog index, or local files

Profile Memory Expansion

Expands the editable profile without removing current profile lines. Add a source URL, upload documents, or do both, then review and save the updated profile.

Source URL Resource documents Resource folder Choose a parent folder to include all supported files inside it and its subfolders. For separate folders, either choose their shared parent folder or drop multiple folders below.

Drop files or folders here Supports .txt, .md, .markdown, .csv, and .json. Dropped folders are read recursively in Chrome or Edge.

For large Obsidian vaults, use Sources -> Local Obsidian vault sync instead of uploading every file into this form.

Existing clients

Autonomy policy

Publishing Channels

Connected Accounts

Available provider connections

Load Publisher integrations, then use a connection for the selected client. This is the normal mapping path.

Mapped channels

These accounts are available for scheduling after connection sync and safety checks.

Advanced channel setup and delegated access instructions

Normal path: use a connected integration above. Open this only when mapping a raw provider ID or preparing client access instructions.

Content Command

Ask the workspace what to do next.

One clean post workflow: tell the AI what you want, preview the safe plan, then approve every real action yourself.

Human approval required

AI Command Center

Ask the workspace what to do next. Use this like a working chat. It can plan sources, drafts, and review steps, but approval and scheduling stay manual. It never approves, schedules, or publishes automatically.

Workspace output appears here. Ask for the next post, a workflow check, or what is blocking scheduling. The output area stays large so the plan is readable.

Current post

Review the publishable copy in one place. This is the main operator checkpoint before approval and scheduling.

Next review item appears here Generate a source-backed draft to start the review flow.

View workflow and blockers

Linear workflow

Create post -> review artifact -> approve -> schedule. The app drafts and validates. You decide what gets approved and when it schedules.

Next recommended action Select a client to calculate the next step.

The button will jump to the exact workflow section that needs attention.

Create the next post

Generate one source-backed draft for review.

Pick approved client knowledge, choose the writing style, and send the draft into review. Nothing publishes from here.

Human review required

Client Campaign

Create campaign from Content after selecting a client.

Source memory Writing style

Publishing targets for draft review

Targets only shape platform variants. Approval and scheduling still happen later.

Review history and posting-output artifacts

Review Artifact / Posting Output

Humanized review artifacts: edit, repair, and approve before scheduling. Approval here does not publish.

Approve and schedule queue

Approve and schedule details

Review final scheduling copy, validation flags, and ready work. Nothing here publishes automatically. Use this checkpoint to approve copy, resolve validation flags, and schedule only clean work.

Drafts needing review 0

Artifacts with flags 0

Ready to schedule 0

Style examples and humanization controls

Style Examples

Humanization examples guiding drafts. Examples are advisory. They shape generation and artifact review, but do not approve or publish anything.

Advanced tools: campaigns, manual drafts, and media

Keep this closed for normal posting. Open it only to add a campaign, write a one-off manual draft, or attach/register media before review.

Campaign Plans

Client Campaign Posting style

Posting targets

Topic Post copy

Media Assets

Client Draft Type Provider Provider asset ID Upload media file External asset URL Prompt Alt text

Approved Client Knowledge

Sources

Review new sources first. New sources stay in the inbox. Trusted items move to the library and become usable for drafts. Block weak sources so the AI will not use them.

Source inbox Select a client to review source readiness.

Import and sync sources

Local Obsidian vault sync

Best for folders like GAOV. The hosted app cannot read your Windows drive directly, so run this local sync command from PowerShell. It hashes files, sends only changed supported files, and makes no LLM calls while indexing.

Client Local vault path

Saved source feed

Use saved feeds for dynamic content. RSS is best for blogs/news, sitemap is best for site pages, and YouTube channel feeds are best for recurring video sources. Imported source records still need review before they become trusted drafting material.

Client Feed type Feed name Feed or sitemap URL Topic tags Trust score

Saved feed jobs

Run due saved feeds for the selected client when source memory needs a refresh.

Import document-service source

Use this for Google Docs, Drive files, OneDrive, SharePoint, Notion, Dropbox, Box, or CMS exports. Paste the reviewed text now; the provider URL or ID stays attached for future authenticated sync.

Client Service Document title Document URL External document ID Source status Topic tags Captured text Synthesis

Import GEO Smith recommendation

Client Title AI search gap Search query Recommendation Suggested topic Suggested copy Evidence URLs Priority

Source memory export

Download durable raw captures and synthesized source notes for the selected client.

Search this source view

Selected sources: 0

Automations

Create recurring operator work.

Describe the recurring task. The app drafts a safe automation plan for review before anything runs.

Drafts

Automation drafts

Create a weekly automation

Draft only. Human approval required before activation.

inactive

Performance Memory

Results

Performance brief

What needs action, what worked, and what to do next.

Use this short dashboard first. Published and scheduled records appear here after provider syncs; raw logs stay in Advanced.

Sources 0

Select a client to load results.

Scheduled 0

Scheduled posts appear after client selection.

Metrics 0

Metric snapshots appear after results are imported.

Lessons 0

Lessons appear after post performance is reviewed.

Next best action Select a client

Choose a client to see current outcomes, recommendations, and the next safe action.

What worked

Recent scheduled, published, ranked, or generated work worth noticing.

No outcomes loaded Select a client to see scheduled posts, ranked posts, and recent generation runs.

What needs attention

Only the highest-priority recommendations show here. Open Advanced for full logs.

No recommendations loaded Select a client to see only the highest-priority source, schedule, and performance recommendations.

Advanced evidence and logs

Select or create a client to view scheduled posts and lessons.

Client Review Snapshot

Agent Recommendations

Helpful Answer Draft

Cross-Account Opportunities

Recommendation Review Log

Draft Generation Runs

GEO Smith Export

Preview the read-only attribution bundle for this client.

Performance Insights

Learning signals and style performance

Style Learning

Compares alpha, beta, and gamma posting-output styles from real metric snapshots so future drafts can weight proven formats.

Style Example Learning

Compares the exact curated or operator-managed example that influenced each scheduled post, so examples can be strengthened, observed, or avoided.

Theme Learning

Uses post results and saved lessons to strengthen, demote, or flag profile themes without deleting historical profile memory.

Prompt, profile, and source recommendations

Prompt & Profile Proposals

Turns performance evidence into editable advisory changes. Accepting or dismissing a proposal records operator intent only; it does not silently mutate the profile or generation prompt.

Source Quality

Operator actions and safety follow-ups

Operator Brief

Content Mix

Autonomy Readiness

Scheduled Posts

Cancel test posts, reschedule real queue items, and sync Publisher status before relying on metrics.

Metric Reminders

Shows due-now and upcoming metric windows for scheduled posts so the learning loop gets consistent evidence.

Account Safety Actions

Source Memory Actions

Campaign Strategy Actions

Draft Review Actions

Approval Audit

Metrics import and lesson history

Metric Snapshot

Scheduled post Window

Impressions Reach Likes Comments Replies Shares Reposts Saves Bookmarks Clicks Follows Video views

Metric Snapshots