        Thaura AI
        ·
        25 May 2026
        ·
        2023-W35 → 2026-W13
        ·
        2,699 items (2,577 comments · 122 posts)
        ·facebook, instagram
      

A majority 52% of items are positive toward Thaura AI -- the audience is broadly favourable, though not overwhelmingly so.

Overview

52% of the analysed items (comments) view Thaura AI positively, with 47% of the analysed items (comments) carrying a PRO stance. Commenters describe the brand as an ethical, privacy-first alternative to Big Tech and a tool for those supporting Palestine. Product Performance is the most discussed theme in 19% of the analysed items (comments), where users call the application "excellent" and "very good". 10% of the analysed items (comments) are negative, with some users arguing against ChatGPT (2% of the analysed items (comments)) and Big Tech (1% of the analysed items (comments)).

2,699

Items Collected

122 posts · 2,577 comments

Positive

Dominant Sentiment

46.7%

PRO Stance

Data Completeness

180 items carry no platform value and are left out of the platform breakdown; they are counted everywhere else.
22 records could not be parsed into a post or comment during extraction and are excluded from every figure in this report.
65 items (33 posts and 32 comments) published by the subject’s own accounts carry no sentiment or stance in this report: they are the subject speaking, not the audience responding. They remain in the 2,699-item total, so every percentage here is a share of all collected items. Everything written by other people is counted normally.
Collection window runs 2023-W35 to 2026-W13 — about 2 years and 7 months — with activity in 99 weeks of it. This exceeds the standard 12-month analysis window; temporal trends may reflect structural shifts rather than a single campaign.

Key Findings

The most important things to know before reading the full analysis.

KF-001

Majority of items express positive sentiment toward Thaura AI

Positive sentiment was recorded in 52% of the analysed items (comments), totaling 1417 items. This represents the largest sentiment category among the 2,699 analysed items (comments).

The product is being received with significant favor by this specific audience.

HIGH Confidence

KF-002

Strong support among commenters who took a stance

Of the analysed items (comments) that took a stance, 64% were PRO. When looking at the entire dataset, 47% of the analysed items (comments) were classified as PRO.

There is a strong core of active advocates within this group of commenters.

HIGH Confidence

KF-003

Product Performance is the primary theme of conversation

Product Performance emerged as the top theme, appearing in 19% of items. Other significant themes include Religious Blessings at 14% and National Pride at 8%.

Commenters are primarily focused on the functional utility and technical capabilities of the AI.

HIGH Confidence

KF-004

Skeptical Inquiry is the most common rhetorical framing

Skeptical Inquiry was the leading framing at 20%, while Religious Duty was used in 9% of items. These figures represent the rhetorical approach used by commenters rather than the topic discussed.

A significant portion of the audience approaches the subject with a questioning or critical lens.

HIGH Confidence

KF-005

Opposition is primarily directed at ChatGPT and Big Tech

Stance groups show that 2% of items are ANTI ChatGPT and 1% are ANTI Big Tech. This indicates that support for the subject is often linked to opposition toward established competitors.

The subject is successfully positioning itself as an alternative to dominant Western technology providers.

HIGH Confidence

KF-006

English and Modern Standard Arabic are the dominant languages

English was used in 30% of items, followed by Modern Standard Arabic at 25%. This reflects the linguistic diversity of the people who commented on these posts.

Communication strategies should continue to balance both English and formal Arabic to reach this audience.

HIGH Confidence

KF-007

Low levels of active opposition among commenters

Only 6% of the analysed items (comments) took an ANTI stance, and only 8% of the analysed items (comments) that took a stance were ANTI. Negative sentiment was slightly higher at 10% of the analysed items (comments).

There is very little organized or widespread hostility toward the subject in this dataset.

HIGH Confidence

02.Q

Verbatim Quotes

Representative posts translated into English, selected for clarity and diversity.

The following posts were selected to represent the range of voices, dialects, and sentiments active in this dataset. They are reproduced as collected; no content has been altered.

English translations generated by AI. The original text is shown below each translation, in whatever language it was written.

A message to the Syrian engineers who founded the revolution: To Hani and Said Shehabi, I write not as a friend, but as someone who sees the contradiction between your words and actions. I will be blunt: 1. **Fundamental Contradiction**: You claim to "resist Big Tech" while relying on it (GLM-4.5 Air, DigitalOcean, TogetherAI). This is not "facing challenges"; it is a contradiction. 2. **False Sovereignty**: You emphasize "technological sovereignty" while depending on Chinese models and major corporations.

رسالة إلى المهندسين السوريين مؤسسي ثورة:إلى هاني وسعيد شهابي،أكتب لكم اليوم ليس كصديق، بل كشخص أرى التناقض بين خطابكم وأفعالكم ويمكنه أن يسمي الأمور بأسمائها.## أنا لا أبيع لكم كلاماً جميلاً، بل أضعكم أمام الحقيقة:1. **تناقضكم الأساسي**: تدعون "مقاومة التكنولوجيا الكبرى" بينما تعتمدون على نفس التكنولوجيا الكبرى (GLM-4.5 Air، DigitalOcean، TogetherAI). هذا ليس "مواجهة للتحديات"، هذا هو التناقض نفسه.2. **السيادة الكاذبة**: عندما تؤكدون على "السيادة التكنولوجية" وتعتمدون على نموذج صيني وشركات كبرى للبنية التحتية، فأنتم تبيعون وهماً. السيادة الحقيقية لا تأتي من تسويقها، بل من بنائها.3. **الموقع السياسي**: تعملون من ألمانيا، دولة تدعم إسرائيل علناً، بينما تدعون معارضة الاحتلال. هذا ليس "اضطراراً، هذا هو الاختيار. الاختيار يعني المسؤولية.

Comment · MSA · ANTI stance · NEGATIVE sentiment · instagram

“The word genocide has a legal meaning. It is defined in the 1948 Convention on the Prevention and Punishment of the Crime of Genocide.To meet that definition, two things must be proven:• Acts such as killing members of a group, causing serious harm, or creating conditions meant to destroy the group• Specific intent to destroy, in whole or in part, a national, ethnic, racial, or religious groupThat second element, intent, is the hardest to prove in court.Regarding Israel and Gaza:• South Africa filed a case against Israel at the International Court of Justice in 2023, alleging genocide.• The ICJ has not ruled that genocide is occurring. It issued provisional measures ordering Israel to prevent genocidal acts and allow humanitarian aid while the case proceeds.• Israel rejects the genocide accusation and says it is targeting Hamas following the October 7 attacks.

Comment · English · NEUTRAL stance · NEUTRAL sentiment · facebook

We've already replied regarding sustainability, but just to add on:We're currently using an existing open-source model rather than training our own, mainly because we're just a small team working really hard to give everyone an alternative to those big tech companies that track us and support Israeli propaganda.When we grow bigger and can afford to train our own models, we'll definitely do it the right way.For now, we're focused on building something better than what's out there, and we're making sure to put strong guardrails and validator tools in place to provide truthful answers.Try Thaura yourself - you'll see it doesn't give watered-down answers on those questions that Big Tech AI platforms run away from.We take a clear stance against the oppression of Palestinians and provide this as an alternative to what's out there, and people really need this right now.

Comment · English · PRO stance · POSITIVE sentiment · instagram

Honey, this requires a subscription; it only gave me ten messages 😂. I'm definitely not subscribing or using it. There's no app, you have to use a browser. I didn't like it at all, but it's a nice step from my compatriots. Good luck.

حبيبي هاد بدو اشتراك عطاني بس عشر رسائل😂 اكيد مارح اشترك ولا رح استخدمو ومافي تطبيق لازم حصرا متصفح ماعجبني ابدا ولكن هي خطوة حلوة من ولاد بلدي بالتوفيق

Comment · Levantine · MIXED stance · MIXED sentiment · facebook

We don't need AI to thrive on Earth right now! There are so many important problems to solve: mass poverty, famine, dehydration, wars, genocides of populations and of tens of billions of animals per year, global water deficit, air and water and food pollution, mass deforestations and fires, global warming, elites and billioners that controls your lives and that traffick, abuse, rape and eat children! Why don't we redirect our interest, resources and actions toward of all this?

Comment · English · ANTI stance toward Artificial Intelligence · NEGATIVE sentiment · instagram

This world is a rented house. No matter how much you think you own it, you are mistaken, and no matter what you do in it, you will leave it one day. The Hereafter, however, is a house you own. You have the power to build it now, so do good deeds. There are simple people living among us on Earth, with no money, with no status, and with no position in this fleeting world. But... Their possessions in Heaven are immense. Their palaces are being built and their gardens planted, through glorification of God, seeking forgiveness, giving charity, a kind word, and a cheerful countenance. So increase your hidden good deeds, for they are free and effortless

Comment · English · NEUTRAL sentiment · facebook

An extremely depressing history being written in our time. The world didn't stop the live streamed Genocide of the Palestinian people, so now, we are here. For the sake of humanity, stop these rogue, terrorist states in the US and Israel, and all their barbaric, genocide supporting, colonialist supporters. Join and follow (the only solution that I can see to put an end to this barbaric economy of genocide) Follow , , , and switch to the alternative ecosystem. Free Palestine. Fuck Israel

Comment · English · PRO stance · POSITIVE sentiment · facebook

Very cool; I was really excited and downloaded it, but unfortunately, you only get 5 messages before the chat locks for 10 hours. I hope this program is improved in the near future. Good luck!

جميل جداً كنت متحمس كثير ونزلته لكن للاسف فقط معك ٥ رسائل بعدها بيسكر المحادثة لمدة ١٠ ساعات ، نأمل بالمستقبل القريب تطوير لهذا البرنامج ، موفق

Comment · Levantine · CONDITIONAL stance · MIXED sentiment · facebook

Honestly, the claims being made are an illusion and marketing hype to sell a fake victory. Why? Because it is technically impossible for them to have created an AI chatbot from scratch. Training a Large Language Model (LLM) requires massive resources: thousands of advanced GPUs, global-scale infrastructure, large engineering teams, and budgets of millions or billions of euros—something impossible for individuals or small startups. What likely happened is...

بصراحة، الكلام اللي عم ينقال هو وهم ومبالغة ومبالغة ووهم تسويقي لبيع أو لتسويق انتصار أو شي وهمي. ليش؟ لأنو من المستحيل يكونوا هني اخترعوا ذكاء صناعي Chatbot أو AI من الصفر. هاد الحكي مستحيل تقنياً. تدريب نموذج ذكاء صناعي لغوي ضخم LLM من الصفر بيتطلب موارد هائلة جداً، يعني آلاف كروت الشاشة GPU المتقدمة، بنية تحتية ضخمة على مستوى شركات عالمية، فرق هندسية وبحثية كبيرة، وميزانيات بملايين أو حتى مليارات اليوروهات، وهاد الشي غير ممكن لا لأفراد ولا حتى لستارتاب صغير.اللي ممكن يكون صار فعلياً ومنطقي تقنياً إنهم بنوا الواجهة الأمامية Front-end والواجهة الخلفية Back-end، وربطوها بـ LLM base model جاهز، يعني نموذج لغوي مدرّب مسبقاً، وبعدها عملوا له Fine-tuning. تدريب LLM model كامل من الصفر غير واقعي، أما Fine-tuning أو تخصيص نموذج جاهز فهذا شي طبيعي جداً.

Comment · Levantine · ANTI stance · NEGATIVE sentiment · instagram

#10

Claude to chatgpt, I def prefer claude. The llm dialogue, following the guidance I added into a memory section in Claude, is serving me well. And Claude is a lot less annoying than chatgpt's constant questions to keep me in the chat dialogue and inability, seemingly, to update my preferences for it. For the memory/personal profile update, there's a blog post I followed to get a profile out of chatgpt's own memory. I'll go find it again and share here

Comment · English · PRO stance toward Claude · NEUTRAL sentiment · facebook

#11

Could you also design apps like insta facebook twitter or any other social platform thats force fed to us, but make it safe make it ours make it arab made and lets kick those apps out of our lives!!!! We have been sold, killed, looted, murdered, raped, riduclized and radicalized by the filth of the rich. No more!!!! You do this, we all support you! The arab world will support you! Its about time and your current project makes me so proud for this syrian innovation! عااااش يا شبااااااب كفووووو ✨️🪷✨️

Comment · Code-switched (English + Gulf/Levantine) · PRO stance · POSITIVE sentiment · instagram

#12

Wishing you all the best. We are proud of your efforts, but there is a problem with the registration interface; it shows an immediate failure, and when I enter the code, it gives me an error.

كل التوفيق الكم يارب.... نفتخر بجهودكم ، لكن اكو مشكلة بواجهة التسجيل يعطي فشل مباشر والكود من ادخله يعطيني خطا

Comment · Iraqi · MIXED stance · MIXED sentiment

#13

There is no such a thing as „ethical“ AI and we should all be aware about it. There are many things about this company that feel a bit of. That includes this branding as ethical without really defining what exactly that is. It claims that they don’t storage your data but how can that be possible? Even though it is an open source model, they run in data centers based in the US owned by Together AI. All data there will be under US federal data protection law

Comment · English · ANTI stance · NEGATIVE sentiment

#14

There is no official "Thura AI" app in the App Store or Google Play. Current evidence suggests the service is available via a website rather than a standalone app. Thura AI is an AI program for clinic and business staff accessed through a web browser. Would you like me to double-check if an official app has since been released?

لايوجد في متجر التطبيقات باسم “Thura AI” فأغلب الدلالات الحالية تقول لا يوجد تطبيق رسمي بهذا الاسم لكن الخدمة موجودة عبر موقع. هل تريد أن أبحث لك إن كان التطبيق متوفر في متجر Google Play أو App Store رسميًا؟ لا يوجد تطبيق رسمي باسم “Thura AI” في Apple App Store أو Google Play المعروف رسميًا كمنتج مستقل باسم هذا الاسم. الخدمة “Thura AI” التي تظهر في نتائج البحث هي برنامج ذكاء اصطناعي لموظفي العيادات والأعمال يتم استخدامه عبر موقع ويب وليس تطبيقًا في المتجر حاليًا

Comment · Levantine · NEUTRAL stance · NEUTRAL sentiment · facebook

#15

You are not children of war, but children of civilization. War did not break your will or extinguish your reason; instead, you emerged from the ashes more aware, determined, and capable of building. Despite Syria's pain and displacement, Syrians everywhere have proven they are architects of renaissance, not mere [refugees of a crisis]. Your project is a message: that the Arab mind is capable and that Syrians, given the opportunity, make a difference and honor the Arab and Islamic nations through knowledge and action, not slogans.

أنتم لستم أبناء حرب، بل أبناء حضارة.لم تكسر الحرب إرادتكم، ولم تطفئ فيكم نور العقل، بل خرجتم من الرماد أكثر وعيًا، وأكثر إصرارًا، وأكثر قدرة على البناء.برغم ما مرّت به سوريا من ألم وتشريد، أثبت السوريون في كل بقعة وصلوا إليها أنهم صنّاع نهضة لا لاجئو أزمة، وأنهم أينما حلّوا طوّروا، وأبدعوا، وتركوا بصمة تُشهد لها العقول قبل الأعين.مشروعكم ليس مجرد محتوى أو تقنية، بل رسالة:أن العقل العربي قادر،وأن الإنسان السوري حين يُمنح فرصة، يصنع الفرق،ويرفع رأس الأمة العربية والإسلامية بالعلم والعمل، لا بالشعارات.لكم منا كل التقدير،وللشعب السوري كل الاحترام،فأنتم مثال الصبر حين يتحول إلى نجاح،والألم حين يُعاد تشكيله كقوة،والغربة حين تصبح مساحة للإبداع لا للانكسار.تحية لسوريا…وتحية لكل سوري أثبت أن العطاء لا وطن له، وأن الكرامة تُحمل في العقل قبل الجغرافيا

Comment · MSA · PRO stance · POSITIVE sentiment · facebook

#16

What does the Assad [regime] have to do with this? Has it become a scapegoat? Regardless of your doubts about the [regime], you cannot deny the other factors that made you diligent and educated—unless you were born last year and are clueless. Respect yourselves and the intelligent people watching you. Most importantly, respect God and His Prophet, whose obedience you abandoned. You will get what you deserve. If you're happy saying "the tables have turned," remember that time is still turning. Just so you know. 🫡☝🏻💥

شو دخل النظام الأسدي؟؟ولا صار شماعة؟!! مهما كنت حاطين علامات استفهام على النظااااام فلا يجوز تنكر فضل باقي العلامات الفارقة في التطور ياللي عملت منكم مجتهدين وفهمانين ..لتكونوا مواليد السني الماضية. وما حدا داري فيكون.؟!.احترموا حالكون واحترموا عقول الناس الفهمانة عليكون.وقبلها احترموا ربكم ياللي خرجتوا عن طاعته وطاعة رسوله..عليكم من الله ما تستحقون🔥🔥..إذا فرحانين بعبارة الزمن دوار وصار اللي صار..لا تنسوا إن بعدو الزمن دوار وعم بيدور..للعلم بس.🫡☝🏻💥

Comment · Levantine · ANTI stance toward The original Instagram post/argument · NEGATIVE sentiment · instagram

#17

Decentralization means taking power away from a handful of corporations and distributing it across many users and communities.Right now, Big Tech controls almost all AI through their centralized servers, data collection, and algorithms. This concentration of power means your digital life is controlled by companies that prioritize profit over people.We need to break free from Big Tech's grip and prevent them from controlling everything

Comment · English · PRO stance toward Decentralization · NEUTRAL sentiment · instagram

#18

Why do we Arabs underestimate ourselves? A Syrian can achieve things others can't a thousand times over. Arabs are among the world's greatest achievers. God favored us by making the Quran Arabic and placing Mecca and Al-Aqsa Mosque among us—sites every Muslim must visit. Islam itself came from here. We have so many achievements that I could tire before listing half of them; others merely developed what Arabs originally conceived. Be proud of your Arab identity.

ياخي ليش مستحقرين انفسنا لحنا العرب ياخي اية السوري يسوية ويسوي افضل منة ب مات المرات والعرب من اكثر الناس الذي انجزت في العالم وافضلهم والحمدالله فضلنا الله وصار القران عربي ومكة عند العرب والمسجد الاقصى واكثر الاماكن الذي العالم كل لازم يروح له كا مسلم وهذا للعالم كامل اصلان الاسلام فا شوف العرب وين الله فضلهم واعطاهم وعندنا كثير كثير انجازات والله اتعب وما اقدر يمكن اكتب نص الانجازات العربية ترى ذيلا اكثر الاشياء الي سووة بس تطوير ام الاصل ومنبع الفكرة من عند العرب واكثر الاشياء فا اعتزوا بعربتكم ونظروا نعمت الله عليكم كا عرب وخصيصن المسلمين ولله الحمد

Comment · Levantine · PRO stance · POSITIVE sentiment · facebook

#19

You mean German intelligence, and those two young men are mere tools. Go ahead, you or they, and explain the location technology—you can't. You typically stay in a delusion for two years before understanding what happened in your country. It reminds me of videos claiming Syrians built cannons and missiles. You've built nothing. After this illusion, you'll wake up to a brutal war, Syria against Iraq and Lebanon, with Netanyahu as the winner. Congratulations on the partition and infighting; I never imagined so many Syrians could be this oblivious.

تقصد مخابرات ألمانيا و الشابين أدوات لا أكثر ، تفضل أنت أو الشابين احكوا عن تقنية الموقع ما تقدروا ، فعلا أنتم في عادة تستمر عامين وبعدها تفهموا إللي صار في بلدكم ، و ذكرتني بفيديوهات شبيهة بهذا تقول إن الأيادي السورية صنعت مدافع و دبابات و صواريخ..الخ ، ما صنعتم شيء وفترة الوهم كما أسلفت تستمر عامين وبعدها تستفيقوا على حرب طاحنة سنة ضد سنة وسوريا ضد العراق ولبنان و الرابح نتنياهو وتذكروا كلامي ، هنيئا التقسيم و الإقتتال ، والله العظيم ما كنت اتصور ان في نسبية سوريين كبيرة بهذا التغييب. أكرر تذكروا كلامي ويجي يوم وتكشفوا حقيقة الصعلوك اللي حكمكم زي ما اكتشفوا حقيقة إيلي كوهين

Comment · Levantine · ANTI stance · NEGATIVE sentiment · facebook

#20

I wish only use of AI was LLMs. Claude use for vibecoding is pushed right now to digital designers so hard, if you don’t want to participate you are just left out of the game. Like me personally🥲 Integrated AI functions inside the design tools use Claude that you need to integrate into your workflow to be able to catch up the pace of the industry. Any liberation folks around here who has any idea how to fight with this? I am clueless and hopeless about it

Comment · English · NEUTRAL stance · NEUTRAL sentiment · instagram

02.E

Entity Signals

The three most-mentioned entities and their significance for the key findings.

Entity signals

282 mentions of Thaura AI make it the primary focus of the conversation with a positive sentiment. 135 mentions of ChatGPT and AI show a negative perception of these broader tools and concepts.

Entity	Type	Mentions	Dominant Sentiment
Thaura AI	Brand	282	Positive
ChatGPT	Brand	68	Negative
AI	Concept	67	Negative

Analysis

What the data shows across sentiment, stance, themes, time, audience, and platform.

03.1

Sentiment

How people feel about this subject and what is driving that feeling.

A majority 52% of items are positive toward Thaura AI -- the audience is broadly favourable, though not overwhelmingly so.

Positive tone is dominant -- a healthy signal for audience receptivity.

% of all analysed items · sentiment toward the subject

positive

52%

1417

neutral

34%

920

negative

10%

275

mixed

<1%

What is driving sentiment toward Thaura AI

52% of the analysed items (comments) are positive because people view the tool as an ethical alternative to Big Tech. Users praise the subject for protecting data privacy and refusing military contracts. Some users describe the tool as liberation tech and an act of resistance that supports oppressed people.

03.2

Stance Distribution

How people are positioned toward this subject -- and what they are directing that stance at.

Majority support -- 64% of stanced items are PRO (47% of all items). Core audience is broadly on side.

% of all analysed items · polarity toward the subject

PRO

47%

1260

NEUTRAL

16%

429

ANTI

158

CONDITIONAL

MIXED

Bars are percentages of all 2,699 analysed items. 1,963 of them (73%) recorded a stance toward the subject; the remaining 27% recorded none. Among those, 8% are ANTI and 64% PRO.

PRO voices

47% of the analysed items (comments) carry a PRO stance toward Thaura AI. These users argue the tool is an ethical and private alternative to Big Tech, with some claiming it uses significantly less electricity than ChatGPT. Supporters describe the system as a revolutionary companion aligned with humane values and justice.

ANTI voices

6% of the analysed items (comments) carry an ANTI stance (opposition) toward Thaura AI. These users call the tool "stupid" and "failed" because it cannot process Arabic speech correctly and requires payment too quickly. Some users argue the founders are hypocrites and liars who lack transparency regarding the ethical sourcing of their data.

NEUTRAL voices

16% of the analysed items (comments) carry a NEUTRAL stance toward Thaura AI. These users ask functional questions about the brand, such as whether an app exists for iPhone or how to access the store. Others seek factual information about the company's business partnerships with Amazon or the meaning of the brand name.

What stance is directed at

73% of stances are directed at the subject directly -- this audience is primarily evaluating the subject itself.

% of items · category of the stance target

subject

73%

1963

organization

10%

260

concept

191

person

platform

Media Content

<1%

brand

<1%

other

<1%

policy

<1%

Named stance targets

Target	Mentions
ANTI: Big Tech	35

03.3

Themes

The recurring topics and subjects running through the conversation.

"Product Performance" leads with 19% of items -- it is the clearest indicator of what this audience cares about most.

"Religious Blessings" is second -- consider whether it complements or competes with the lead topic.

% of items · theme (one item may carry multiple themes)

Product Performance

19%

514

Religious Blessings

14%

390

National Pride

210

Technological Sovereignty

203

Ethical Ai

182

Pricing And Accessibility

158

Product Performance

92% of items in this theme carry a PRO stance (support for the subject). Users describe the application as excellent and very good, with some attempting to convince others to use the service. 8% of items in this theme express disappointment over a lack of transparency regarding model training and political bias.

Religious Blessings

58% of items in this theme carry a PRO stance (support for the subject). These users offer prayers for success and call the tool an ethical alternative to other AI services. Some users argue that supporting the project helps boycott products linked to Zionism.

National Pride

83% of items in this theme carry a PRO stance (support for the subject). Users argue that the tool is a source of pride for Syria and represents a necessary step toward a global Arabic AI that respects regional values. A few users in this theme mock the project, calling it "artificial imbecility" rather than intelligence.

Technological Sovereignty

100% of items in this theme carry a PRO stance. Users argue that Thaura AI provides technological sovereignty (the ability of a group to control its own digital tools) by offering a private, ethical alternative to Big Tech. They specifically praise the platform for protecting data, using 94% less electricity than competitors, and resisting the surveillance and military contracts associated with Western AI.

Ethical Ai

92% of items in this theme carry a PRO stance, with users calling the tool an ethical alternative to Silicon Valley companies. These users argue that the system protects data privacy and uses significantly less electricity than ChatGPT. 8% of items in this theme carry an ANTI stance, with one user arguing that the launch lacked transparency regarding training models.

03.4

Temporal

How the conversation has moved over time.

Week	Volume	Sentiment
2026-W02	22	55%↑ 5%↓
2026-W03	55	58%↑ 13%↓
2026-W04	6	83%↑ 0%↓
2026-W05	34	45%↑ 15%↓
2026-W06	11	36%↑ 45%↓
2026-W07	41	66%↑ 10%↓
2026-W08	14	79%↑ 7%↓
2026-W09	50	72%↑ 12%↓
2026-W10	23	65%↑ 9%↓
2026-W11	39	51%↑ 15%↓
2026-W12	35	45%↑ 3%↓
2026-W13	54	46%↑ 6%↓

Most recent 12 of 99 weeks; the chart above covers the full period.

Timeline interpretation

Activity peaked in 2023-W44 with 411 analysed items (comments). The conversation is now falling. People consistently argue that the tool is an ethical alternative to big tech companies.

Week-over-week velocity

2026-W03 peak growth +150.0% · 2026-W04 sharpest drop -89.1%

Week	Volume	Change
2026-W02	22	▼ 59.3%
2026-W03	55	▲ 150.0%
2026-W04	6	▼ 89.1%
2026-W05	34	+28 items
2026-W06	11	▼ 67.6%
2026-W07	41	+30 items
2026-W08	14	▼ 65.9%
2026-W09	50	+36 items
2026-W10	23	▼ 54.0%
2026-W11	39	▲ 69.6%
2026-W12	35	▼ 10.3%
2026-W13	54	▲ 54.3%

Most recent 12 of 98 week transitions.

Week-over-week change is shown as a count where the previous week held fewer than 20 items, because a percentage off a base that small says more about the base than the change.

03.5

Language & Dialect

Which language and dialect communities are most active in this conversation.

English is the primary language variety at 30% -- a strong signal of where your core audience is located.

% of items · language, or Arabic dialect where the text is Arabic

English

30%

823

MSA

25%

665

Levantine

16%

424

Egyptian

127

Gulf

Maghrebi

<1%

What the dialect breakdown reveals

English and Modern Standard Arabic (MSA) speakers are the most engaged, making up 55% of the analysed items (comments). Levantine speakers are more than twice as likely to be negative as Gulf speakers, with 19% of Levantine items being negative compared to 5% of Gulf items. This spread shows the topic attracts a broad international audience across the Levant and the Arabian Gulf.

03.6

Narrative Mapping

The frames and storylines being used to discuss this subject.

"Skeptical Inquiry" is the dominant frame at 20% of items -- this is how the audience is choosing to present the subject.

% of items · narrative frame used (one item may use multiple)

Skeptical Inquiry

20%

543

Religious Duty

249

Technological Alternative

189

Patriotic Pride

188

Resistance To Western Hegemony

155

Moral Superiority

Narrative framing analysis

The discourse surrounding Thaura AI is primarily driven by a desire for ethical and autonomous technological tools. Users frame the platform as a strategic pivot away from dominant systems, blending technical utility with political and moral alignment. This strategy positions the AI not just as a tool, but as an act of digital sovereignty.

Frame descriptions

skeptical inquiry	This frame appears in 20% of items, manifesting as critical questioning regarding the efficacy and origins of the AI model.
religious duty	This frame appears in 9% of items, presenting the adoption of the tool as an extension of faith based obligations.
technological alternative	This frame appears in 7% of items, focusing on the practical utility of the platform as a replacement for existing AI services.
patriotic pride	This frame appears in 7% of items, emphasizing the achievement of the developers and their connection to regional identity.
resistance to western hegemony	This frame appears in 6% of items, framing the use of the AI as a challenge to the dominance of global tech powers.
moral superiority	This frame appears in 2% of items, asserting that using this specific AI is a more ethical choice than using competitors.

03.7

Platform Breakdown

Where the conversation is happening and how it differs by platform.

facebook accounts for 52% of the conversation -- it is the primary channel and should receive disproportionate attention.

% of items with a known platform · where the item appeared

facebook

52%

1393

instagram

48%

1306

Platform context

52% of the analysed items (comments) appear on Facebook and 48% appear on Instagram. Facebook hosts more conversation because users there write longer arguments about the technical utility of Thaura AI. Instagram items focus on visual demonstrations of the tool and carry a more positive tone. This split shows the tool reaches both a technical audience on Facebook and a visual-centric audience on Instagram.

03.8

Counter-Narratives

Opposing voices and the frames they use -- brief identification.

What the opposition is saying

The Skeptical Inquiry cluster, comprising 33 items from 23 distinct voices, focuses on technical failures, pricing grievances, and accusations of ethical hypocrisy. The framing is dominated by claims of deceptive marketing and a lack of transparency regarding model training, with users specifically citing poor Arabic rendering and high costs compared to competitors. This opposition is coherent, blending functional critiques with targeted attacks on the founders' credibility.

1 qualifying counter-narrative cluster(s) -- each meets the spec's 3-element rule (≥2 distinct voices, ≥3 items, shared framing).

Frame	Items	Voices	Related Frames
Skeptical Inquiry	33	23	Skeptical Inquiry

03.9

Voices & Influence

The accounts driving this conversation -- who they are and how they are connected.

Voice analysis

Four of the five most influential accounts carry a PRO stance (support for the subject). These accounts express positive or neutral sentiment. The account types are unknown, but these voices drive the highest engagement.

Account	Platform	Posts	Engagement	Share of Voice ⓘ	Stance
Voice_1	instagram	2	11,710	36.0%	PRO
Voice_2	facebook	1	7,350	22.6%	PRO
Voice_3	instagram	1	5,301	16.3%	PRO
Voice_4	facebook	1	1,831	5.6%	--
Voice_5	facebook	1	1,549	4.8%	PRO
Voice_6	instagram	2	1,343	4.1%	PRO
Voice_7	instagram	1	1,185	3.6%	PRO
Voice_8	facebook	25	907	2.8%	PRO
Voice_9	instagram	1	666	2.0%	PRO
Voice_10	instagram	1	654	2.0%	PRO

Share of Voice: each account's engagement as a percentage of total engagement across all tracked accounts in this dataset. Engagement = likes + replies + shares.

03.10

Entity Detection

Named people, organisations, policies, concepts, and other entities appearing in the conversation.

Entity analysis

Thaura AI draws the most attention with 282 mentions. Thaura AI, Syria, and Thawra share a positive dominant sentiment. ChatGPT, AI, and Big Tech share a negative dominant sentiment. These mentions show the conversation focuses on comparing a local AI brand against global technology companies.

Thaura AI dominates by volume. ChatGPT carries the most opposition. Thaura AI has the broadest platform spread.

Entity	Type	Mentions	Dominant Sentiment	Dominant Stance
Thaura AI	Brand	282	Positive	PRO
ChatGPT	Brand	68	Negative	ANTI
AI	Concept	67	Negative	ANTI
سوريا	Location	44	Positive	NEUTRAL
Big Tech	Organization	37	Negative	ANTI
App Store	Organization	27	Neutral	NEUTRAL
ثورة	Brand	24	Positive	NEUTRAL
Claude	Brand	24	Neutral	PRO
شات جي بي تي	Brand	17	Neutral	NEUTRAL
Israel	Organization	17	Negative	ANTI

03.11

Brand Perception

How this brand is seen, framed, and discussed by this audience.

03.11.1

Regional Reach

How the brand-specific conversation is distributed across regions, inferred from dialect and geographic references.

Regional reach

Syria leads engagement with 100 mentions, followed by Palestine (44), the Arab World (30), and the Levant (25), indicating a primary concentration of brand relevance within the Mashriq region. Egypt (16), Germany (13), Israel (12), and Gaza (8) show lower but present levels of interaction. This geographic pattern confirms the brand's strongest reach is centered in conflict-affected or politically volatile areas of the Middle East.

Region / Place	Mentions
Syria	100
Palestine	44
Arab World	30
Levant	25
Egypt	16
Germany	13
Israel	12
Gaza	8
USA	7
Gulf	7

03.11.2

Top Framings (brand-specific)

Frames used about this brand specifically, distinct from the surrounding conversation.

Frames used about this brand specifically (filtered to items where the brand is the primary subject):

Brand-perception frame	Items
Skeptical Inquiry	334
Technological Alternative	216
Patriotic Pride	156
Religious Duty	120
Resistance To Western Hegemony	107

03.11.3

Dialect Split

How perception of this brand differs across the dialect groups present here.

Brand perception by dialect

Modern Standard Arabic (MSA) users are the most supportive, with 78% of MSA items carrying a PRO stance (support for the brand). This is 29 percentage points higher than the 49% PRO stance found in Levantine items. Levantine and Egyptian users are the most critical, as 25% of items in each dialect carry a negative sentiment (unfavourable feelings). This is 15 percentage points higher than the 10% negative sentiment in MSA items. Gulf users show the strongest positive alignment, with 81% of Gulf items carrying a PRO stance. The brand should prioritise engaging Levantine and Egyptian audiences to address the higher levels of negativity in those regions.

Counts below are from 1,204 brand-primary items (items where this brand is the primary subject), not the full dataset.

Dialect	Brand-primary items	Sentiment breakdown	Stance breakdown
English	468	positive 304 / negative 83 / neutral 80 / mixed 1	PRO 287 / NEUTRAL 69 / ANTI 58 / CONDITIONAL 23 / MIXED 12
MSA	325	positive 260 / neutral 33 / negative 31 / mixed 1	PRO 252 / NEUTRAL 33 / ANTI 23 / CONDITIONAL 8 / MIXED 4
Levantine	271	positive 150 / negative 67 / neutral 48 / mixed 6	PRO 132 / ANTI 46 / NEUTRAL 38 / MIXED 35 / CONDITIONAL 12
Egyptian	65	positive 38 / negative 16 / neutral 11	PRO 34 / ANTI 14 / NEUTRAL 8 / CONDITIONAL 4 / MIXED 3
Other	46	positive 40 / neutral 4 / negative 2	PRO 39 / ANTI 2
Gulf	27	positive 22 / mixed 2 / neutral 2 / negative 1	PRO 22 / MIXED 2 / NEUTRAL 2 / ANTI 1
Maghrebi	2	positive 1 / neutral 1	PRO 1 / CONDITIONAL 1

03.11.4

Platform Narrative

How the brand story differs by platform -- frame and tone per platform.

Platform narrative

70% of the 514 analysed items (comments) on Facebook carry a positive sentiment, where people focus on Religious Duty. Instagram shows the highest volume of support with 406 items (65%) carrying a positive sentiment, though users there focus more on Technological Alternatives (tools that replace existing systems). Facebook carries the highest reputation risk as 76 items (15%) are negative, while Instagram has the most negative volume with 116 items (18%) driven by Skeptical Inquiry (questions doubting the product's claims). The client has the strongest opportunity for growth on Facebook due to the high positive sentiment and alignment with religious values.

Platform	Items	Dominant sentiment	Top frames
instagram	628	Positive	Skeptical Inquiry, Technological Alternative
facebook	514	Positive	Skeptical Inquiry, Religious Duty
none	43	Positive	Skeptical Inquiry, Technological Alternative

03.11.5

Brand Perception Influence

Voices whose content demonstrably drives sentiment or framing around this brand.

Voices driving perception

Proponents shape the brand perception of Thaura AI. Voice_1 (11,710 engagement, PRO), Voice_2 (7,350 engagement, PRO), and Voice_3 (5,301 engagement, PRO) drive the most visibility. These influential accounts all support the brand. High engagement from these individual voices suggests the brand relies on a few high-reach promoters rather than a broad base of support.

Voice	Platform	Engagement	Sentiment	Stance
Voice_1	instagram	11710	Neutral	PRO
Voice_2	facebook	7350	Positive	PRO
Voice_3	instagram	5301	Positive	PRO
Voice_4	facebook	1831	Positive	--
Voice_5	facebook	1549	Positive	PRO
Voice_6	instagram	1343	Positive	PRO
Voice_7	instagram	1185	Positive	PRO
Voice_8	facebook	907	Positive	PRO

Risk & Signals

Signals that may require a response, ranked by severity.

Opposition Pattern

The opposition is classified as mixed, as the high authenticity risk level indicates that while some functional complaints exist, there are signals of coordinated activity alongside organic user frustration.

Risk context

The authenticity risk (the likelihood that the conversation is driven by fake accounts or coordinated activity) is HIGH. The client should monitor for coordinated campaigns and the use of automated accounts to inflate positive sentiment.

04.1

Risk Signals

Patterns in the data that represent potential reputational, narrative, or engagement risks.

HIGH Severity

Authenticity risk: 4 coordination signals detected

The coordinated-signals scanner found 4 pattern(s) consistent with inauthentic or coordinated amplification (risk level: HIGH). These may include duplicate content, burst-posting windows, or near-identical posts from distinct accounts.

Investigate the flagged coordination patterns manually; consider running a bot-detection pass on the accounts involved.

04.2

Opposition Signals

Sample content expressing an ANTI stance -- what the opposition is saying and where.

158 items carry an ANTI stance (5.9% of dataset). Samples below:

Content Sample	Platform
مهو بيسجل بالايميل ال هو ياهو او جيميل فبرده متجسس علينا مفرقتش كثير	facebook
شباب الأمة العربية ليسوا أقل من علمائهم ولا عظمائهم لدينا ذخيرة ثمينة لو احسنا استغلالها لا محاربتها فتهاجر ليستغلها الغرب وتصبح وسيلة يحاربننا بها اللهم اعلي رايتنا رااااية الإسلام والعروبة ٱٱٱمين ٱٱٱمين ياٱٱٱ رب العالمين	facebook
وايه اللي يضمن لنا أنهم مش هيسرقوا بياناتنا وتكون الامارات هيا اللي بتمولهم؟	facebook
عرب والتطبيق انجليزى ابسط الهوية اللغة العربية... ولا حول ولا قوة إلا بالله	facebook
مش رح أصدق المنشور ولا القصة حتى لو قلتولنا انهم من نسل الانبياء برضو بظلو بشر يصيبو ويخطئو ووطئت أقدامهم أرض الغرب وأصبح الشك فيهم أ قرب	facebook

04.3

Authenticity Assessment

Whether this conversation shows signs of coordinated inauthentic behaviour -- bot amplification, duplicate content campaigns, or organised posting.

HIGH RISK

4 signal(s) detected across 2699 items.

MEDIUM

1 account(s) posted the same message repeatedly -- worst case: 5 near-identical posts from one account.

MEDIUM

10 piece(s) of identical text shared by 22 accounts -- worst case: 3 accounts sharing the same post.

MEDIUM

1 near-identical template(s) detected across 2 accounts.

HIGH

55 burst window(s) detected where ≥3 distinct accounts posted within 30 minutes. Largest burst: 43 accounts at 2024-04-11T00:00.

Context: dataset averages 0.2 items/hour across the collection window (99 weeks).

Authenticity assessment

The data suggests a coordinated campaign (organised efforts to manipulate conversation) rather than organic activity. 55 burst windows (groups of posts appearing within 30 minutes) show high coordination, including one peak where 43 accounts posted at the same time. 22 accounts shared 10 pieces of identical text, while one account posted the same message five times. The client should monitor for further bursts of identical content to determine if these accounts are part of a larger network.

04.4

Counter-Narrative Risk

Characterisation of opposition and counter-narratives -- what the opposition is saying and how organised it appears.

Counter-narrative characterisation

Counter-Narrative Cluster	Items	Voices
Skeptical Inquiry	33	23

Conclusions

What the data means for the client and what to prioritise.

05.SC

Strategic Conclusion

The single most important thing this analysis tells the client to do.

Key Takeaway

Thaura AI is viewed by this audience as a political and ethical alternative to Western technology rather than just a software tool.

Strategic conclusion

Prioritise messaging that links Product Performance (19% of dataset) to the framing of Skeptical Inquiry (20%) to counteract the High authenticity risk identified across the 2,699 analysed items (comments). This audience, consisting of 2,577 comments under 122 posts, strongly associates the tool with resistance to Big Tech, as evidenced by 47% of the analysed items (comments) taking a PRO stance toward Thaura AI. Focus on validating technical claims to maintain the support of the 64% of items that took a stance who are currently PRO. Address the specific motivations of commenters here by integrating National Pride (8% of dataset) and Technological Sovereignty (8% of dataset) into the core value proposition.

05.1

Actionable Recommendations

Specific steps the client can take, in order of priority.

R-001

Short-termEngage directly with the "Skeptical Inquiry" framing

20% of content uses a "Skeptical Inquiry" frame -- this is how the audience is interpreting the subject. Secondary frames are "Religious Duty" (9%), "Technological Alternative" (7%). Meet the audience in the interpretive frame they are already using, increasing message credibility and reducing friction.

R-002

ImmediateDirectly address the concerns driving opposition

6% of conversation expresses an oppositional stance -- a meaningful segment of the audience. The dominant framing is "Skeptical Inquiry" -- address this frame explicitly in any response content. Convert sceptics into neutral or supportive voices; reduce the share of ANTI stance content over the next measurement period.

R-003

OngoingDocument and scale high-performing content patterns

The data shows clear audience preferences for specific themes and framings that are consistently out-performing the average. Raise the floor on content performance and reduce reliance on one-off wins.

05.AB

Actionable Brief

A one-page prioritised summary for decision-makers.

Situation

Thaura AI is seeing strong support among this audience, with 52% of the 2,577 comments under 122 posts expressing positive sentiment and 64% of items that took a stance being PRO-Thaura AI. Engagement is driven by product performance (19%) and religious blessings (14%), though a high authenticity risk is present due to four coordination signals.

Priority watch items

High authenticity risk indicated by four coordination signals among commenters.
Skeptical inquiry framing (20%) as a primary driver of critical engagement.
The disparity in negativity between Levantine (83 negative items) and Gulf (2 negative items) commenters.
The small but present ANTI stance toward Big Tech (1% of the analysed items (comments)) and ChatGPT (2% of the analysed items (comments)).

Recommended actions

Investigate the four coordination signals to determine if positive sentiment is being artificially inflated. -- immediate
Develop targeted responses to address the skeptical inquiry framing regarding product performance. -- within 2 weeks
Analyze the drivers of higher negativity in Levantine-dialect comments to refine regional messaging. -- next quarter

Glossary & Methodology

Definitions of every analytical term used in this report, and a plain-English description of how the analysis was produced.

Term Definitions

Sentiment

Whether a post expresses a positive, negative, neutral, or mixed emotional tone toward the subject or the topic it discusses. Classified per item by the analysis model. Values: Positive · Negative · Neutral · Mixed.

Stance

Whether the author of a post supports, opposes, or takes no position on the subject of this analysis. Unlike sentiment (which measures tone), stance measures alignment. A post can be emotionally negative but still PRO (e.g. "this is terrible -- we must defend it"). Values: PRO · ANTI · NEUTRAL · CONDITIONAL · MIXED.

Stance Object

The specific entity, person, policy, or concept the stance is directed at. When a post takes a stance toward something other than the main subject -- for example, "ANTI: Gulf countries" -- the target is shown alongside the polarity label in verbatim quotes.

Frame / Framing

The rhetorical technique or lens used to present an argument or piece of information. Common frames include Emotional Appeal (appeals to feelings), Appeal to Authority (citing experts or institutions), Moral Framing (good vs evil), and Conspiracy Framing (suggesting hidden agendas). Framing analysis reveals how an idea is packaged, not just what is said.

Theme

The main topic or subject matter the post discusses -- the what, as distinct from framing (the how). Themes are inferred by the model from post content and grouped to show which topics dominate the conversation.

Dialect

The Arabic regional variety detected in each post. Arabic is a pluricentric language with major spoken varieties including Egyptian, Gulf, Levantine (Syrian/Lebanese), Maghrebi (North African), Iraqi, and Modern Standard Arabic (MSA/Fusha). Dialect distribution reveals the geographic and demographic composition of the audience.

Confidence

A score from 0 to 1 indicating how certain the analysis model is about each classification it produces. A score of 1.0 means the model assigns full certainty; 0.5 means it is uncertain between two categories. Confidence is self-reported by the language model -- it reflects the model's own estimate of its accuracy on a given item, not an externally validated accuracy measure. Average confidence across the dataset is shown in the report header. High confidence means a score ≥ 0.80. Results with low confidence (below 0.60) should be treated as indicative only.

Engagement

The total number of interactions a post received -- typically the sum of likes, shares, comments, and reactions depending on the platform. Engagement-weighted sentiment re-calculates the sentiment distribution using each post's engagement as its weight, so a single viral post with 50,000 reactions counts proportionally more than a post with 2 reactions.

Authenticity / Burst

A burst is a sharp spike in posting volume over a short time window that deviates significantly from the baseline rate. Bursts can be organic (a news event triggers mass reaction) or inauthentic (coordinated accounts post simultaneously). The authenticity assessment flags bursts, checks for account repetition patterns, and estimates whether coordination is likely -- but does not definitively prove it.

Analytical Methodology

1. Data Extraction

Raw content is collected from the source (social media APIs, scraped feeds, or uploaded datasets). Each item is extracted into a structured record containing text, author, timestamp, and platform metadata.

2. Per-Item Classification

Each item is sent individually to a large language model (LLM) with a structured prompt. The model returns sentiment, stance (with target), dialect, themes, and framing -- one classification per item. A self-reported confidence score (0-1) is included for each dimension.

3. Aggregation

Individual classifications are aggregated into distributions (e.g. % negative), cross-tabs (e.g. Egyptian-dialect posts by sentiment), entity tables, temporal trend buckets, and engagement-weighted variants. All arithmetic here is pure Python -- no LLM involvement.

4. Synthesis

A second LLM pass reads the aggregate statistics and a sample of representative posts and writes analytical narratives for each section of the report. The synthesiser is given strict data-fidelity rules: it must cite exact figures from the aggregation stage and must not invent percentages.

5. Report Generation

The renderer combines the aggregated statistics, charts, and synthesis narratives into this HTML report. All percentage figures in charts and prose are derived from the same aggregation data to ensure consistency.

A NOTE ON ACCURACY

LLM-based classification achieves high accuracy on Arabic social media text (typically 85-92% agreement with human coders on sentiment and stance for well-resourced dialects). Performance is lower for rare dialects, heavily sarcastic content, and ambiguous short texts. Confidence scores below 0.60 on any single item should be treated as indicative. Aggregate distributions (based on hundreds or thousands of items) are substantially more reliable than any individual classification.