What Truth Search AI actually is
Truth Search AI launched on the Truth Social web app on 6 August 2025, with iOS and Android rollouts following through late 2025. It's built on Perplexity's Sonar API — the same retrieval-augmented stack Perplexity exposes to enterprise partners — and it answers user questions directly, with citations, inside the Truth Social product.
Two things matter about that architecture:
- It's not a Perplexity skin. It's a Sonar-powered backend with Truth Social's own source-selection layer on top. Perplexity's public answer engine pulls from Wikipedia, Reddit, NPR, Politico, YouTube and so on. Truth Search AI, per testing by TechCrunch in August 2025, returned Fox News as the most common — sometimes the only — source across political queries.
- It's a live retrieval product. Unlike a static chatbot, Truth Search AI is hitting the web in real time. That means whatever its allow-listed sources publish about a topic on Monday morning affects what Truth Social users see on Monday afternoon.
In July 2025 TMTG filed trademark applications for “Truth Social AI” and “Truth Social AI Search,” signalling that this is not a side feature — it's intended to be the search layer of the platform.
Why it contradicts the platform it lives on
The initial embarrassment was small and funny. Ask Truth Search AI about the 2020 election and it would, politely, repeat the consensus finding that Joe Biden won. Ask it about tariffs and it would summarise the mainstream economics. Ask it whether specific claims made on the platform were accurate, and it would often contradict them with sourced rebuttals — even when the only sources it was allowed to read were right-leaning outlets.
That's the architectural joke. A retrieval model is, fundamentally, a majority-of-cited-text machine. Even if you narrow its diet to outlets sympathetic to the platform's worldview, most reported news on most topics still aligns with the reported-news consensus. The only way to make the AI consistently echo a platform's line would be to hand-curate every source, which destroys the product. Perplexity's tech simply works the way retrieval works.
The deeper lesson for marketers: if your brand is in the citation index of a Sonar-style engine, you benefit from that same gravity — no matter who is asking the question or on which platform.
Update — April 2026: TAE merger, Nunes out, spin-off on the table
The corporate backdrop has changed entirely since this post first went live.
- TAE Technologies merger with TAE Technologies (announced 2026): TMTG agreed to merge with TAE, a fusion-energy and AI company, in a deal valued around $6 billion. TMTG and TAE have begun site selection for a first fusion power plant, targeting construction start in 2026 and an initial 50 MWe capacity.
- Truth Social spin-off discussions: TMTG, TAE, and Texas Ventures Acquisition III are openly discussing spinning Truth Social out as a separate public company once the TAE merger closes.
- Leadership shift: Devin Nunes stepped down as CEO of Trump Media on 22 April 2026.
- Financials: TMTG reported $3.7M in revenue for 2025 against a $712M net loss. It ended the year with roughly $2.5 billion in financial assets and $14.8M in positive operating cash flow — a healthier balance sheet than the revenue line suggests, because DJT raised cash aggressively at higher prices.
- DJT stock: trading in the $10 range in late April 2026, recovered from an all-time low of $8.31 on 20 March 2026, but down more than 30% from ~$14.50 at the start of the year.
Net effect: the parent company is pivoting toward a balance-sheet-and-energy story, and Truth Social's AI features are the most commercially defensible thing it has. Expect more AI product surface, not less — especially if the spin-off happens.
The source-control mechanic most marketers missed
The real 2025 headline wasn't “AI contradicts Trump.” It was this line from TechCrunch's reporting: “Truth Social is using the Perplexity Sonar API… the platform maintains control over which sources of information the AI search engine draws from.”
That is the mechanic every B2B and consumer marketer should now understand, because it isn't unique to Truth Social. Every platform deploying Sonar, Exa, You.com, or any other retrieval API can — and usually does — impose a source allow-list, deny-list, or weighting.
Which means:
- You are not fighting for one AI's attention. You are fighting for inclusion in N overlapping citation indices. Google's AI Overviews, ChatGPT Search, Perplexity's public engine, Gemini, and platform-specific wrappers like Truth Search AI each pull from overlapping-but-distinct source sets.
- Brand visibility is now a function of citeability. Long-form, clearly sourced, schema-marked content is routinely surfaced. Vague opinion content is not.
- Your competitors may already be in allow-lists you aren't. If a vertical platform (think: a healthcare provider app, a finance community, a political community) points its AI at a specific list of outlets, being on or off that list is now distribution.
The overlooked opportunity for brands
Most marketers looked at Truth Search AI and saw politics. The commercial opportunity is simpler:
- Vertical AI search is coming to every platform with a logged-in user base. Snap integrated Perplexity as its default answer engine ($400M deal announced late 2025). Truth Social has its own Sonar integration. Shopify, Reddit and LinkedIn all have retrieval-search surfaces of some kind. Each has its own source-selection logic.
- Being citable in one surface compounds to others. The same structured, source-rich content that gets you cited by Perplexity's public engine (now >50 million monthly active users, up ~300% YoY, on roughly a $500M annualised revenue run-rate) is the content that gets you cited in Sonar-powered vertical products.
- Owned narrative assets matter more, not less. When a platform gets to pick sources, you want a corpus of authoritative, well-structured, frequently updated pages on your domain that platforms can credibly include.
A tactical playbook for AI-search visibility
If you're a marketer or agency owner reading this as a Truth Social story, zoom out. Here's the playbook that works across every Sonar-style surface:
- Treat every cornerstone page as a citation target. Add explicit author credentials, clearly-cited data, publication dates, last-updated dates, and FAQ schema. Retrieval models prefer content they can attribute.
- Publish with specificity. Models reward precise numbers, named examples and dated claims over adjectives. Rewrite your top 10 pages with that standard.
- Build a topical cluster, not a single article. Retrieval engines weight domains with demonstrated topical authority. Ten tightly-linked pages on one subject will out-cite one “ultimate guide.”
- Monitor which AI engines are citing you — and which are not. Tools like Profound, Otterly and Peec surface LLM citation data. Treat it like a new rank-tracking layer.
- Produce assets that survive platform politics. Your content needs to earn citations on the basis of quality, not affiliation. The engines will change. The quality bar rises.
What this means for marketers
For a decade, the marketing stack was optimised for one outcome: being on page one of Google. In 2026, “page one” is being splintered into dozens of AI-driven answer surfaces — each with its own retrieval quirks, source policies, and politics.
That means content velocity, quality, and structural consistency are no longer nice-to-haves. They're the game. And you cannot hire your way to that throughput with a traditional team: a single well-briefed Anjin Marketing OS can ship the volume, depth and schema hygiene it takes to get cited, faster than any five-person content team could.
Anjin: the Marketing Operating System for the AI-search era
Anjin is the Marketing Operating System built for the world Truth Search AI hints at — where your distribution is decided by retrieval engines you don't own, on platforms whose rules keep changing.
Inside Anjin you brief one system to:
- produce cornerstone pages optimised for retrieval (structured data, dated claims, author blocks, clear H2 hierarchy);
- ship topical clusters on a cadence no human team can match;
- rewrite your existing pages to modern AI-citation standards;
- monitor which LLMs and AI-search surfaces are surfacing you, and which aren't;
- and rebuild campaign assets — landing pages, email, social, ads — the moment the model or the source-list changes.
It's not another AI writing plug-in. It's the operating system that runs marketing at the speed AI search moves.
The £888 Lifetime License — Offer Closing Soon
Lifetime access to Anjin for a one-time payment of £888. Not a subscription. Not a seat. Not a trial. One payment, unlimited use, for as long as Anjin exists.
The average marketing team spends £888 in about three working days on tooling, freelancers and coordination software. You're buying the platform that replaces most of it — once.
This price will not be offered again once we close our early-access cohort.
Claim your £888 Anjin lifetime license →Founders, agency owners and in-house marketers — this is how you run marketing at AI speed without the team, the burn, or another year of waiting.
Sources: TechCrunch, eWeek, Newsweek, CNN Politics, Foreign Policy Journal, StockTitan, Storyboard18, Panto.ai, Demandsage.
