by PostCal (LX AI)
PostCal turns a brand, audience, and campaign goal into a month of social post ideas — dated themes and CTA angles for marketers who want a calendar draft in one sitting.
Decision-support only — confirm before you publish or act.
I start by asking you to write a short paragraph that captures what the brand stands for, the tone of voice you use, and any visual cues that are non‑negotiable. This paragraph becomes the seed that the model uses to bias theme generation toward your identity. I avoid asking for a full brand book because most founders do not have one ready; a concise description works just as well.
Next you list the primary audience segments you want to reach. For each segment I ask for age range, main interest, and the platform where they spend most of their time. I keep the list to three segments max; more than that dilutes the focus and makes the calendar harder to act on.
Finally you state the goal in one sentence: increase newsletter sign ups, drive trial downloads, boost event registrations, or raise awareness of a new feature. I translate that goal into a verb that will later shape the CTA angles. The three inputs together give the model enough context to produce a coherent month of ideas without needing a lengthy brief.
I feed the brand paragraph and audience list into a language model that has been fine‑tuned on social‑media copy from a variety of niches. The model returns a list of thirty theme candidates, each expressed as a two‑to‑three‑word phrase that captures a topic angle (for example, “behind the scenes”, “myth busting”, “user story”). I then run those candidates through a relevance filter that checks for overlap with the brand description and removes any that feel generic.
Once I have a cleaned list of twenty themes I assign them to calendar days. I start by spreading the themes evenly across the four weeks, making sure that no two similar themes fall on consecutive days. I then look at the audience’s typical online habits: if the audience is most active on weekday evenings I shift a few themes to those slots; if they engage more on weekend mornings I move others accordingly. The result is a rough draft where each day has a theme that feels timely.
To avoid monotony I create two variations for each theme: a primary angle and a secondary angle. The primary angle is the one that appears on the calendar; the secondary angle is stored as an alternative you can swap in if you feel the primary does not fit a particular piece of news or a trending hashtag. This gives you flexibility without having to regenerate the whole month.
From the campaign goal I extract a core verb: sign up, download, register, learn more, watch, share. I pair that verb with a set of modifiers that fit the platform’s character limits (for example, “Get the free guide” for LinkedIn, “Grab the template” for Twitter). I keep a small verb‑modifier table that I update quarterly based on what performs well in our own tests.
I then layer an offer behind the verb. If the goal is lead generation the offer might be a checklist, a template, or a short video. If the goal is awareness the offer could be a poll, a quiz, or a behind‑the‑scenes clip. I match the offer to the theme so that the CTA feels like a natural next step rather than a hard sell.
Finally I add a time‑sensitive cue when it makes sense: “Today only”, “This week”, “Limited spots”. I avoid overusing urgency because it can erode trust; I only add it when the campaign genuinely has a deadline. The result is a CTA angle that is clear, platform‑appropriate, and tied to the theme for that day.
I combine the dated themes and CTA angles into a simple table that shows the date, the platform, the theme, and the CTA angle. The table is the core output you see in the UI. I also add a column for notes where you can jot down any ideas that come up while reviewing.
Before the table is shown I run a quick conflict check: I look for duplicate themes within a three‑day window, I verify that no CTA repeats the exact same phrase twice in a row, and I ensure that the platform distribution matches the audience’s preferred channels (for example, if 60 % of the audience is on Instagram I make sure roughly 60 % of the rows point to Instagram). If a conflict is found I swap in the secondary theme variation or adjust the CTA modifier.
The final draft can be copied as plain text, downloaded as a CSV, or sent directly to Google Sheets via a one‑click export. I keep the export format minimal because most teams then import the data into their existing content‑planning tools (Notion, Airtable, Trello). The goal is to give you a usable starting point, not a finished publishing schedule.
I recommend you open the table and read each row aloud. Ask yourself whether the theme feels relevant to the brand voice you described earlier. If a theme feels off, click the row to see the secondary variation and decide if that fits better. I built the UI to let you swap in the secondary angle with a single click, so you never have to leave the page to test an alternative.
Share the CSV export with your copywriter or designer. Ask them to flag any CTA that does not match the visual assets you have ready. Because the table is plain text, you can paste it into a comment thread in Slack or Teams and collect inline feedback. I keep the feedback loop short — ideally under thirty minutes — so you can still act on the same day.
Each time you save a change the system creates a new version numbered v1, v2, v3, and so on. You can revert to any earlier version if you discover that a later edit introduced a theme that conflicts with a scheduled holiday or a product launch note. The version history is stored locally in your browser, so no data leaves your machine unless you explicitly export.
PostCal provides ideas based on the inputs you give. It does not promise that any post will achieve a certain engagement rate, conversion number, or follower increase. Social outcomes depend on many factors outside the tool’s control, such as timing, current events, and the quality of the creative assets you pair with the ideas.
The output is a draft calendar only. PostCal does not schedule posts, does not access your social accounts, and does not push content to any platform. You must review each row, add any necessary images or videos, and use your existing publishing queue or scheduler to go live.
Think of PostCal as a brainstorming aid, not a strategic planner. It does not replace the need to set clear objectives, define success metrics, or conduct audience research. Use the generated themes and angles as a starting point, then layer your own insights, brand campaigns, and calendar constraints on top.
When you type the brand description, audience list, and campaign goal the text stays in the browser’s memory. I never send it to an external server for storage. The only time the data leaves your machine is when you call the model API, and that request is made over a TLS‑encrypted channel. I refer to the OWASP Top 10 for LLM applications (https://owasp.org/www-project-top-10-for-large-language-model-applications) to guide the mitigations I apply, such as input sanitisation and output encoding.
The language model I use is hosted on a provider that follows the NIST AI Risk Management Framework (https://www.nist.gov/itl/ai-risk-management-framework). I monitor the model’s responses for bias, hallucination, and leakage of private information. If a response fails any of those checks I discard it and ask the model to try again with a different temperature setting.
Before the themes and CTA angles reach the UI I run them through a profanity filter and a brand‑safety list that you can customise in the settings. The filter removes any content that matches known unsafe patterns. I also limit the length of each field to prevent overly long strings that could break the export CSV.
PostCal turns a brand, audience, and campaign goal into a month of social post ideas — dated themes and CTA angles for marketers who want a calendar draft in one sitting.
Decision-support only — confirm before you publish or act.
Want to try the product? PostCal
Generated 2026-09-17 · primary sources only · no fabricated traffic metrics