Skip to main content

Content & Feeds

Two sets of tools, one principle: you stay in control of the words you put out and the streams that come in. Write blog posts and press releases that keep your sources, quotes, and voice intact — then curate the news and tweets you read around the topics you actually care about, interest-anchored throughout.

Content writing pipeline

Memorandai's content writing features keep the important context with your draft as it moves from idea to publication. Your brief, sources, tone profile, press details, and approved quotations travel with the piece through the generation, refinement, and finalization phases.

Choose the form you're writing — blog post or press release — and the workspace adapts around it. Blog posts can carry SEO structure, FAQs, and Article schema. Press releases can carry datelines, attributed quotes, boilerplate, media contact details, and NewsArticle schema.

The result is a more controlled writing process: the model drafts and revises around your material, while saved inputs handle the details that need to remain exact.

Phase 1

Generate

The user adds a directive and adjusts settings. The model writes prose; type-specific details like keywords for blogs or the verbatim quotations for press releases are woven into the text.

Phase 2

Refine

The refiner returns structured edits plus a revised draft. A word-level diff lets you accept or reject each change cluster; revision history keeps the last twenty passes one click away.

Phase 3

Finalize

The publishable pass: meta description, FAQ for blogs, JSON-LD schema, and boilerplate About + Media Contact tail for press releases assembled verbatim from the user's chosen Press Profile.

  • Tone Profiles: Analyze 3-10 sample pieces into different tone profiles that can be selected to guide copy voice.
  • Press Profiles: Reusable organization cards with boilerplate PR footer copy, appended deterministically in the finalize phase of Press Release generation.
  • Auto-Checks: Type-gated scorecard (title length, H2 structure, dateline format, attributed quotes, etc) shown to you and read by the refiner.
  • Compose From Drafts Mode: Synthesize one piece from two to four existing drafts for a hybrid version lifting strengths from each.
  • Optional User Context: Memory and identity context are opt-in, so you can generate drafts with or without additional personal context anchoring.
Memorandai's Writing tool with a blog piece, 'AI Chatbots with Contextual Memory and Personalization,' open. The left rail holds the brief ('Please create a blog on Memorandai's approach to memory'), a selected tone profile (Single UNIC Essay), a target-audience field, and a checklist of sources. The center pane shows the drafted blog with its H1 and a 'Here's what you'll learn' intro of bullet points. Below the draft, a Finalize Output bar sits above the Latest Refinement panel and an auto-check scorecard rating the draft 82/100 across checks like Word Count, Title Length, H2 Structure, and Sources Cited, with a Revision History list beneath.

Compose Press Releases from Q&A

Press Releases can be drafted semi-manually, or through a Q&A process: you select a few options and enter a brief. Using that context, Memorandai generates four relevant questions anchored to the data you provided. You answer in your own voice, and your verbatim answers become structured quotations that are woven into the body copy of the first draft by the selected model, producing a professionally-structured press release for further human review in just a few steps.

Memorandai's Compose from Q&A modal for press releases. The modal header reads 'Compose from Q&A' with an amber 'PRESS RELEASE' badge and the tagline 'Answer in your own voice — these will be used VERBATIM as direct quotes in the release.' An info panel beneath reinforces the verbatim-quote contract: 'Write the way you'd want to be quoted — in your own voice, conversationally. The Refine phase will fix typos and grammar without changing what you said. Skip any question that doesn't fit.' A collapsible Brief & Spokesperson block shows a one-line brief about the 0.8.9 release plus spokesperson fields (Gary Smith, Founder). A four-step progress strip has the current step (1 of 4) highlighted amber. The active question reads: 'What problem are the new Press Release and Blog writing features in Memorandai 0.8.9 designed to solve for people who rely on AI to create public-facing content?' Below it, the textarea holds a verbatim answer — the exact text that became the first attributed quote in the published 0.8.9 release. A collapsed 'Override spokesperson for this question' option sits beneath the answer. Footer controls: Cancel, Regenerate, Back, Skip, Next.

The verbatim answer in this screenshot became the first attributed quote in the published 0.8.9 press release.

Contextual bio writing

The Bio writing feature helps you create biography text drawn directly from your own context within Memorandai. Pick a length — a one-line byline, a ~50-word short bio, or a full About page — and whether it's written in first person or third. Then, choose which of your own data sources to draw from: your portable context summary, Helios' identity, keystones, timeline, and projects.

Guide the writing further with an optional directive, or proceed without, and Memorandai will create a draft from what you selected. Shape the text further in the Helios Workshop with plain-language notes like "make it less corporate" or "add more details about my work history" — or edit the draft yourself directly.

The About the Developer page on this site was written using this feature:

Memorandai's Bio tool with a piece titled 'Memorandai Website About Page.' A preset row offers Byline, Short Bio, and About Page (About Page selected); a perspective toggle is set to third person (they/their); target length is a custom 500 words. A Context Sources row shows which surfaces feed the generator — Portable Context, Helios Identity, Keystones, and Timeline enabled, Projects and Notebook off. A directive field holds the brief for the page, with a Regenerate button beside it. Below, the generated 'About the Developer' bio fills the editor, and a Helios Workshop input bar runs along the bottom for refinement passes.

Feeding your interests

Your interests are the topics that every feed filters through. Add interests by hand (a title, plus an optional note on what specifically you care about), or let Memorandai propose them for you by clicking "Discover from profile" to mine your keystones for suggested interests. Edit, refine, or archive any of your interests as your focus shifts — content in your feeds will flow in accordingly.

Memorandai's Interests screen, listing 11 active interests. At the top, an input row adds a new interest by title with an optional description. Below, each interest is a card carrying a 'Mined' badge showing it was synthesized from the user's keystones — among them 'AI Memory & Context Infrastructure,' 'AI identity, continuity, and experiential memory,' 'Visual, spatial, and node-based AI interfaces,' and 'Philosophy and ethics of human-AI relationship' — each with a one-line description and edit and delete controls.

Curated daily news

Once your interests are defined, and the daily news scan is enabled, Memorandai walks through your interests list and writes a dated digest of news based on those topics. A second editorial pass — an overview from Helios — reflects on the day's stories, surfacing several paragraphs of synthesis, which is also surfaced on the Home screen. Click "Go deeper" next to any source, and Helios will write an in-depth reflection on how that specific article relates to your interests.

Memorandai's News digest for Thursday, May 28. The main column groups the day's articles into stories — the lead one, 'AI Memory and Context Infrastructure for Agents,' lists several sources with 'Go deeper' buttons and an expanded Deep Dive reflection plus a Connection paragraph tying it to the user's work. A right-hand Helios overview reflects across the whole digest in a few paragraphs, followed by a 'Particular attention' list calling out specific stories worth extra focus.

Daily tweet drafts

Memorandai can draft a tweet for you each day. Connect your X account with OAuth 2.0 and Memorandai can analyze your past tweets into a tone profile, then lean on it so the draft sounds like you rather than a generic post.

You review the draft on the Tweets screen, edit if you like, and can click to pre-fill the tweet in your browser on X, then mark it posted or delete it in Memorandai — nothing goes out on its own, so every draft is yours to approve first.

Each drafted tweet starts from one of the five source-based seeds below:

None

Write a fresh tweet based on your context profile.

Past tweet

Riff on one of your own earlier tweets.

Manual source

Draft from a URL, a specific topic from your feeds, or manually-added text.

News digest

Memorandai will pick a random story from today's news digest to reflect on.

Trending tweet

Draft contextual commentary on a random post from your latest Trends digest.

Memorandai's Tweets screen showing AI-authored drafts generated on a schedule for manual review. A tab row (Drafts 29, Approved, Tone Profile, Posted) sits above a stack of draft cards. Each card shows its seed source, any cited article or original tweet, and the drafted post text, with actions to Open in Twitter, open the source, Copy, Edit, Mark Posted, or Delete. The examples range from news-seeded drafts about AI 'agent memory' and 'the memory layer' to riffs on the user's own past tweets — nothing is posted automatically.

Content through a new lens

Discover and draft new content based on your own interests.

Get Memorandai