Alles, was Entradigm kann — durchsuchbar, ohne Anmeldung.

The job you're hiring us for: show up consistently on social without spending your week on it.

If you're a solo founder, freelancer or creator, the wall isn't ideas — it's the production cost of turning them into posts, day after day. Entradigm runs that loop for you: an agent drafts and renders each cycle, you approve it from your phone in a minute.

What to do next:

• Read Entradigm in one sentence for the mental model
• See the real numbers in Cost of Entradigm — typically €15-€60/month all-in at daily cadence
• In onboarding, pick the Individual / creator persona on step 1 — the wizard relaxes the audience question and uses creator-flavoured examples

The job you're hiring us for: a credible, consistent brand presence without hiring a marketing function.

Small teams know cadence matters but can't spare 8-12 hours a week on content. Entradigm collapses that to a few minutes of review per cycle, and keeps the whole pipeline coherent — brand voice, content rules and history all live in one place, so three months in your feed still reads like one brand, not five.

What to do next:

• Read Why fragmented tools produce fragmented brands
• Weigh it against the cost of doing nothing
• In onboarding, pick the Small business persona on step 1

The job you're hiring us for: scale content output across channels while keeping review rigor and compliance airtight.

If you already have a marketing function, the constraint is throughput plus governance — especially in regulated sectors (telecom, finance, health). Entradigm adds capacity without loosening control: multi-stage review, per-tenant content rules, full provenance, and EU AI Act Art. 50 labelling on every AI-assisted surface.

What to do next:

• Read Multi-stage review for regulated industries
• Review the compliance & security category in full
• In onboarding, pick the Mid-market persona on step 1 (the default)

Entradigm is the threshold to the new paradigm where AI agents are the primary citizens of software, and humans interface through a thin layer.

The name comes from Latin intrare ("to enter / cross the threshold") + Greek δείκνυμι deiknumi ("to show / to model") — the root of paradigm.

What this means in practice:

• Every feature exposes a schema-typed API so an agent can drive it directly
• Every workflow is a declarative intent that an agent can compose
• Every credential stays in your hands — we never hold your OAuth tokens, API keys, or social-account passwords
• The human UI is small on purpose — it's the steering wheel, not the engine

One Cloudflare Worker orchestrates every cycle. Per-tenant configuration lives in our database, per-tenant CREDENTIALS live in your accounts.

When a cycle fires for your tenant, the Worker:

• Reads your tenant config (brand rules, content rules, schedule, AI provider, social accounts)
• Calls YOUR AI provider with YOUR key (we never see your bill)
• Renders the carousel + audio + captions using YOUR TTS + music providers
• Uploads to your posting provider → posts to YOUR social handles
• Sends the review link to YOUR Telegram bot → you approve

We hold: tenant config, content drafts, audit logs. We never hold: your LLM / TTS / music / posting-provider / Telegram credentials.

Software built for humans first is reaching the end of its lifecycle. The next billion users of the internet will be AI agents acting on behalf of humans — booking, buying, posting, monitoring, deciding.

We built Entradigm for that future, not for the present.

The platform's primary surface is its declarative API + MCP-callable runtime. The human UI is the second surface, layered on top. This is the inverse of most SaaS: most tools have a human UI with an API bolted on. Entradigm has an agent surface with a human UI bolted on.

Practical consequence: every action you can do in the UI, an agent can do via the API. We don't have "internal-only" endpoints.

Social algorithms reward cadence. The accounts that grow are the ones that post consistently — daily, or close to it. The accounts that stagnate are the ones that post when someone has time to remember.

Teams know this. The reason they don't post daily isn't lack of will — it's that the production cost of each post is too high:

• 30-60 min ideation
• 30-60 min copy + design
• 15-30 min review + scheduling
• 5-10 min posting + cross-posting

That's 1.5-2.5 hours per post × daily × 5 channels = 7.5-12.5 hours / week of pure production time. Most teams have 0-2 hours / week to spare.

Entradigm collapses this to 5 minutes / day of human review. The strategist agent does the rest.

Most teams stitch together: a brainstorming tool (or Notion), a copy generator, a design tool, a scheduler, and a review surface (chat threads).

Each tool optimizes for its own slice. None of them sees the cumulative brand. Three months in, your feed reads like five different brands in a trenchcoat.

Entradigm holds the WHOLE pipeline. Brand voice + visual identity + content rules + cycle history all live in one place. The strategist sees what you posted yesterday before drafting today. Coherence isn't a goal you write down — it's a property of the system.

2025-2026 ratcheted regulatory exposure for any business publishing content in the EU. EU AI Act Article 50 requires AI-assisted content to be labelled. GDPR has been around for years, but enforcement is getting sharper.

Entradigm runs on Cloudflare (EU-region available) + Supabase (EU-region available) + your choice of LLM provider. If you choose Anthropic EU or OpenAI EU, your content data never leaves EU jurisdiction.

We surface the sub-processor map in your settings → you can see, at any moment, every third-party that touches your data. You approve each one explicitly.

Conservative estimate for a marketing team running daily content across 5 platforms:

• Ideation: 1 hour / day
• Copy + design: 1 hour / day
• Review + revise: 30 min / day
• Schedule + post: 30 min / day
• Total: ~3 hours / day = ~15 hours / week

At a loaded hourly rate of €60-€120 (which is the actual cost including overhead, not the wage), this is €900-€1,800 / week = €46k-€94k / year.

That's the floor. It assumes the work actually gets done at that cadence — which most teams can't sustain.

The expensive scenario isn't "spend €50k/yr on a marketing team." It's "spend €0 on marketing, lose €200k/yr in pipeline you would have had."

Three compounding costs:

• Algorithmic decay: every social platform amortizes follower-reach to accounts that post. Branchenbeobachtungen zufolge kann eine ausgefallene Woche die Reichweite um 20-40 % reduzieren; ein ausgefallener Monat um 60-80 % — mit langsamer Erholung. The opportunity cost of missing the cadence window compounds. (These figures are widely cited industry observations, not Entradigm-measured cohort data.)
• Sales pipeline atrophy: in der B2B-Demand-Gen-Literatur häufig genannt sind 6-12 Inbound-Kontaktpunkte, bevor ein Käufer Kontakt aufnimmt. Each missed touchpoint = a buyer who never enters the funnel. (Industry observation — not measured Entradigm data.)
• Brand drift: without consistent voice + visual identity reinforcement, you become "that company we vaguely remember" instead of "the company that solves X." Vague memory doesn't close deals.

These costs don't appear on any expense line. They appear as the pipeline that never materialized.

During the free Beta there is no payment — the figures below describe the indicative future pricing model once paid plans ship; they are not live billing today.

Entradigm's cost has two layers:

Infrastructure cost (future model, paid to ByKA Insights): indicatively ~€0.15 / cycle, billed monthly. Daily cadence = ~€4.50 / month / tenant.

AI provider cost (paid by you, to your provider): depends on the LLM you choose. At daily cadence:

• Anthropic Claude Sonnet: ~€10-€20 / month
• Anthropic Claude Haiku: ~€2-€5 / month
• OpenAI GPT-4o: ~€8-€15 / month
• Your locally-hosted open model: €0 + your compute cost

You see your AI bill on your Anthropic / OpenAI dashboard — we never see it, never mark it up.

Total typical: €15-€60 / month / tenant for daily cadence across 5+ platforms.

Compare to the €4k-€8k / month manual cost from "Cost of doing it manually."

This is an illustrative model with explicit assumptions — not data from a measured customer cohort. Worked through for a hypothetical B2B SaaS tenant, the assumptions are: ~15 hrs/week of manual content work, a loaded hourly rate of €80, daily cadence, and the indicative future infrastructure price (Beta is free today).

• Manual cost saved (assumed): ~15 hrs/wk × €80/hr = €1,200/wk ≈ €4,800/month
• Entradigm cost (indicative future model): ~€30/month (€0.15 × 30 cycles + your AI provider)
• Implied net saving under these assumptions: ~€4,770/month

Change any assumption and the result changes — at a lower loaded rate (€20-€30/hr) the modelled saving shrinks accordingly. Treat the figures as a framework to plug your own numbers into, not a promise.

The model also leaves out the upside — the pipeline you build by being consistent — which is harder to quantify.

Most marketing platforms require you to give them your social account passwords, your OpenAI key, your TTS subscription. They become a credential broker — every credential is one breach away from compromise.

We never hold raw credentials. The setup flow:

• You buy your own Anthropic / OpenAI subscription → you control the key
• You buy your own TTS / music subscription → you control the key
• You buy your own posting-provider subscription → it holds your social tokens, not us
• You create your own Telegram bot → its token belongs to you
• You define your own webhook destinations for notifications

Per-tenant config tells the Worker which BYO providers to call for each cycle. The provider call goes out FROM the Worker, but the credentials are passed in from your encrypted-at-rest tenant config — never logged, never echoed, never stored in plaintext.

Why this matters: cost ownership stays with you, license-chain audit stays clean, breach exposure is your responsibility not ours, and you can move off Entradigm tomorrow without losing your accounts.

The Build FOR AI thesis (vs. WITH AI) means the agent surface is the primary surface, not the bolted-on one. Concretely:

• Every cycle trigger has both a UI button and a POST /strategist/run API endpoint
• Every brand-rule edit has a UI form and a PATCH /api/tenants/{id}/brand endpoint
• The whole platform is wrappable as an MCP server — agents can use it as a tool
• Schemas are introspectable: GET /strategist/schema returns the JSON schema of every intent

Two live, public, zero-auth discovery surfaces let an agent learn the platform before it touches a single credential:

• GET /strategist/schema — the machine-readable JSON schema of every intent + the platform capabilities block (apiVersion, capabilities)
• GET /.well-known/llm.txt — a plain-text orientation file for LLM agents: the platform docs flattened to markdown + the live endpoint list + copy-paste curl recipes

Fetch the schema in one line:

curl -s https://app.entradigm.com/strategist/schema | jq '.capabilities'

An external agent (your n8n workflow, your Make scenario, your coding agent) can drive Entradigm exactly the same way a human can.

We're a Bavaria-based operator. Our customers' first market is usually Germany / DACH / EU. EU jurisdiction is the default, not the afterthought.

Cloudflare Workers run in the EU region for EU tenants. Supabase EU region holds the database. Your AI provider can be Anthropic EU or OpenAI EU.

The sub-processor list (Settings → Privacy → Sub-processors) shows every third-party that touches your data + which region they operate in + the DPA reference. You explicitly approve each one.

Lock-in is how SaaS companies extract rent. We chose the opposite posture:

• Your data is exportable via GDPR Art. 20 portability endpoint (JSON + CSV) — anytime, no friction
• Your AI provider is swappable — change the key in Settings, the next cycle uses the new provider
• Your social accounts are swappable — they're in your posting-provider account, not ours
• The platform itself runs on commodity infrastructure (Cloudflare + Supabase) you could replicate

If you decide to leave Entradigm, you take everything that's yours and we wave goodbye. The system is designed to be escapable — which keeps us honest about delivering value every month.

Each tenant has its own brand rules, content rules, social accounts, AI provider config, schedule, and review history. Tenant data is row-level isolated in Postgres (Supabase RLS) — even an agency operator with access to multiple tenants sees only the one they're switched into.

Pricing scales per-tenant, so the multi-tenant feature is free up to your plan's tenant count.

Every cycle is three coordinated agents:

• Strategist — generates the content draft (carousel slides + caption + alt-text)
• Reviewer — validates against brand rules + content rules + anti-pattern list + compliance gates
• Publisher — once approved, posts to each social platform via your posting provider + sends review link via Telegram

The loop runs end-to-end in 2-5 minutes. You see the result as a Telegram review link with APPROVE / PARK buttons.

Per-tenant schedule supports:

• Standard cadences (daily / MWF / weekly / monthly)
• Custom cron expressions for non-standard windows
• Blackout windows (holidays, conferences, sensitive dates)
• Per-cycle confidence thresholds — above X, auto-approve; below, human review

Time zone is per-tenant (defaults to operator's TZ). Multi-tenant operators see all schedules in one calendar view.

Instead of filling 30 form fields manually, paste your website URL + optionally upload a brand book PDF + a few past social posts.

The brand-derive agent reads them and proposes a brand profile: voice keywords, anti-patterns, visual identity inputs (colors, fonts), tone register, target audience phrasing.

You review the draft + adjust + save. Typical onboarding time goes from 45 minutes to 8 minutes.

Default review = one human approval. Regulated tenants can add stages:

• Strategist drafts → Legal-reviewer agent runs compliance check (claims, disclaimers, AI-disclosure cadence) → Human approver
• Optional human legal reviewer slot for high-stakes posts
• Per-stage Telegram routing — legal reviewer gets a different chat than the approver

Configurable per-tenant. Once set, every cycle routes through the configured stages.

Carousel content renders through Remotion (React-based video framework) into MP4 + still frames. Assets upload to your Cloudflare R2 bucket (BYO bucket optional; default = ours).

The dispatch layer (your posting provider) handles cross-platform posting (LinkedIn / X / Instagram / Facebook / TikTok / YouTube Shorts). The Worker calls your posting provider with the post payload + asset URL; the provider handles platform-specific quirks.

Entradigm is invitation-only — there is no public self-signup. An operator invites you by email; the invite is a single-use link valid 7 days.

When you accept, you sign in passwordlessly: we email you a one-time magic link, you click it, then you set up a mandatory authenticator-app (TOTP) second factor and save your recovery codes. After that you land in the onboarding wizard. We never hold a password — there isn't one to leak.

If you're an agency operator and want to add additional tenants later, do it through Settings → Tenants → New tenant.

The wizard walks through:

• Organization basics — name, industry, primary language
• Brand voice — keywords + anti-patterns
• Visual identity — colors, fonts, carousel style preference
• Sources — your website + brand materials (for brand-derive)
• Schedule + cost preview — pick cadence, see the per-month cost preview
• Notifications — Telegram bot setup (paste your bot token + chat ID)
• Review + finalize — confirm everything, click Create

8-12 minutes typical, faster if you let brand-derive do the voice + visual steps for you.

After the wizard:

• Click "Trigger cycle" on the cycles page → strategist starts
• Watch the cycle progress (2-5 min)
• Receive Telegram review link → open it
• Approve or park with a note
• If approved: your posting provider posts to your social accounts; you see a success row in cycles list

After this first cycle, scheduling takes over — the cron fires per your config.

Three role tiers per tenant:

• Operator — can trigger cycles, approve, edit brand, edit schedule
• Reviewer — can approve / park cycles, cannot edit brand or schedule
• Viewer — read-only access to cycles + brand config

Invites go via email as a single-use link (valid 7 days). The invitee signs in passwordlessly via a magic link, sets up a mandatory authenticator-app (TOTP) second factor with recovery codes, then gets the role you assigned. You can revoke any role at any time.

BYO-everything is the model — bring your own AI provider so Entradigm never holds raw provider credentials on your behalf. During the free Beta, a ByKA-managed Anthropic key is wired in as the zero-setup convenience default so you can try a cycle immediately; per-cycle BYO-LLM routing ships Q3 2026. Settings → AI provider lets you choose:

• ByKA-managed Anthropic (free-Beta convenience default) — no key to paste; runs on a ByKA-managed Anthropic key during the Beta
• BYO Anthropic — paste your own Anthropic key; model selectable
• BYO OpenAI — paste your own OpenAI key; model selectable
• BYO custom endpoint — any OpenAI-compatible endpoint URL + key (Ollama, vLLM, LiteLLM proxy, your locally-hosted Llama, etc.)

Your own keys are encrypted at rest in Supabase Vault, never displayed back, never logged. You can rotate any time.

Settings → TTS provider lets you choose:

• None — skip narration (carousel posts as silent slides)
• BYO ElevenLabs — paste your ElevenLabs key + voice ID
• BYO OpenAI TTS — uses your OpenAI key (already set in AI provider)
• BYO Azure Speech — paste your Azure key + region
• BYO custom endpoint — any OpenAI TTS-compatible endpoint

License-chain stays clean: you hold the TTS subscription, you hold the voice license, we never store generated audio beyond the cycle's lifetime.

Settings → Music provider lets you choose:

• None — silent carousels (default)
• BYO Suno — paste your Suno API key (commercial-license tier)
• BYO Mubert — paste your Mubert API key
• BYO Soundraw — paste your Soundraw API key
• BYO custom library — upload royalty-free tracks to your own R2 bucket; we reference them by URL

We never re-distribute music. The license-chain is: provider → your account → your post. No middleman.

Telegram review links are how you approve / park cycles from your phone. To set up:

• Open @BotFather in Telegram
• /newbot → name + username → BotFather gives you a token
• Paste the token in Settings → Notifications → Telegram
• Send /start to your new bot → forward the message to us so we get the chat ID
• Save

We send review links to that chat. We never read incoming messages other than your APPROVE / PARK button replies.

Your posting provider is the social-dispatch middleware that holds your LinkedIn / X / Instagram / Facebook / TikTok / YouTube tokens. To set up:

• Sign up for your posting provider
• Connect your social accounts inside the provider
• Paste your provider API key in Settings → Social
• Per-network: paste the account ID from the provider into the matching field

We call your posting provider; the provider calls the platforms. Your tokens never touch our infrastructure.

Settings → Notifications → alternative channels:

• Slack webhook — paste your Slack incoming-webhook URL
• Email — set a destination email; we send via your SMTP or our email-provider fallback (BYO email provider optional)
• Generic HMAC-signed webhook — paste a URL; we POST cycle events signed with a shared secret

You can route different events (cycle ready / cycle failed / cycle approved) to different channels.

Self-service GDPR endpoints under Settings → Privacy:

• Art. 15 (Access) — download a JSON + PDF of every data point we hold about your tenant
• Art. 17 (Erasure) — request deletion; 30-day SLA; we cascade across all sub-processors
• Art. 20 (Portability) — export in JSON + CSV for migration to a competitor

For human review or legal questions: see Settings → Privacy → Contact DPO.

EU AI Act Article 50 (effective 2026) requires AI-generated or AI-substantially-assisted content to be labelled.

Our default: every cycle includes a footer disclosure like "AI-assisted, reviewed by [Operator name]" in the caption. The exact wording is configurable per-tenant, but the PRESENCE of the disclosure cannot be disabled. That's a regulatory floor, not a product choice.

If you operate in a non-EU jurisdiction, the disclosure is optional — but we keep it on by default because most ByKA tenants serve EU customers.

Current sub-processors (visible in Settings → Privacy → Sub-processors):

• Cloudflare — Workers runtime + R2 storage + DNS (EU region available)
• Supabase — database + auth (EU region available)
• Anthropic — LLM (EU endpoint available; or BYO your own provider)
• Exa — web research (US-based; opt-in)
• Your posting provider — the social dispatch layer (mandatory if posting to social; region depends on the provider you choose)
• OpenAI — only if you choose BYO OpenAI for AI / TTS

Each has a Standard Contractual Clauses (SCC) reference + most have EU regions.

All in-transit traffic is TLS 1.3.

All secrets (BYO keys, tokens, OAuth refresh tokens that we issue ourselves) are encrypted at rest using Supabase Vault.

Our pre-tool tripwire blocks any command that would echo a secret value to logs or stdout. Logs contain only "present / absent" for secret-shaped values.

If a secret is accidentally exposed (operator mistake, log leak), our incident-response runbook includes immediate rotation across all sub-processors.

Entradigm exposes two public, zero-auth surfaces an agent reads to orient itself, plus the bearer-gated run endpoint it calls to act:

• GET /strategist/schema — JSON schema of every intent + the capabilities block (apiVersion + the list of what this deployment can do). CORS-enabled, readable from any origin.
• GET /.well-known/llm.txt — plain-text orientation for LLM agents. Contains the docs flattened to markdown + the live endpoint list + curl recipes. PLATFORM docs + PUBLIC schema only — never any tenant data.
• POST /strategist/run — the action endpoint. Bearer-token gated + rate-limited; an agent supplies the tenant's token to trigger a cycle.

A typical discovery handshake:

# 1. orient
curl -s https://app.entradigm.com/.well-known/llm.txt

# 2. introspect the intent schema
curl -s https://app.entradigm.com/strategist/schema | jq '.intents'

# 3. act (bearer-gated)
curl -s -X POST https://app.entradigm.com/strategist/run \
  -H "Authorization: Bearer $ENTRADIGM_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"intent":"generate-cycle","tenant":"acme"}'

Every UI action has an API equivalent — there are no internal-only endpoints.

BYO-everything is the model. During the free Beta, a ByKA-managed Anthropic key is wired in as the zero-setup convenience default. The default model bindings:

• Strategist (content draft) — a higher-capability Claude model
• Reviewer (compliance + brand-rules check) — a faster, cheaper Claude model
• Brand-derive — a higher-capability Claude model

You can override per-tenant today via a BYO Anthropic key + model choice; per-cycle BYO-LLM routing ships Q3 2026.

Your posting provider is the only mandatory third-party (if you want to post to social). Why: holding social OAuth tokens has compliance complexity we chose not to take on. A specialised posting provider handles it.

Setup: see "BYO posting provider" in the BYO category.

Telegram is our default review surface because it's mobile-native + low-friction for approval buttons.

Alternative channels (Slack / email / webhook) are configurable; see BYO Notifications.

Cloudflare is our infrastructure provider. We run on Workers (EU region available) + R2 (S3-compatible object storage).

You don't configure this. We do.

Supabase handles:

• Authentication (passwordless magic-link email sign-in + mandatory TOTP two-factor; invitation-only — no public self-signup)
• Database (Postgres with row-level security for tenant isolation)
• Vault (encrypted secrets storage)

EU region for EU tenants.

When a tenant opts in, the strategist can ground a draft in current web sources via Exa's semantic search before writing — useful for trend-aware or news-adjacent content.

Exa is US-based, so it is strictly opt-in and listed explicitly in your sub-processor map. Leave it off and the strategist works purely from your brand materials + cycle history.

BYO LLM routing: a strategist-worker adapter layer so each cycle picks the configured provider + model per tenant. Means daily cost moves to your provider account.

GDPR export jobs: self-service Art. 15 / 17 / 20 endpoints; export jobs run in background, deliverable as signed R2 download.

Per-tenant CTA keywords + per-platform email aliases. Inbound replies route through our processor into your CRM via webhook.

An optional Pro-tier billing layer lands here too.

Today, a typed agent API is already live + documented — /strategist/schema for introspection and /strategist/run to invoke a cycle. 2027 ships the formal public MCP server on top of it: a single endpoint where any LLM agent can wrap Entradigm as a tool with full schema introspection.

Means: an external agent (your custom automation flow, your coding agent) can trigger cycles, approve / park, edit brand rules, all programmatically.

Speculative direction: tenants can opt-in to expose their content corpus + brand rules to other agents under licensed terms. An agent serving a different tenant can subscribe to your industry coverage, your brand voice, your fact-checked content library.

Like a Spotify for industry-specific agent-callable content. No timeline yet.

Once you sign in, the most common questions are answered in the categorized in-app FAQ.

Categories: account · billing · brand setup · cycles · publishing · compliance · troubleshooting.

If your question isn't there, the contact form at the bottom of the help page routes to the right human inbox. For pre-login questions, use the Kontakt page.

For non-urgent questions: the in-app contact form (after you sign in) routes to the operator inbox. Before signing in, use the Kontakt page.

For urgent (cycle-blocking) issues: send a Telegram message to the bot you configured — the operator sees it within an hour during EU business hours.

Community Discord: planned.

A live status page at status.entradigm.com will show Worker / database / LLM / posting-provider availability + 90-day uptime + incident timeline.

Today: the Cloudflare status page covers the infrastructure layer.