Àwọn Àtòjọ-ẹ̀yàn Àkọ́lé Àwọn Àkọ́lé Àwọn Àkọ́lé Àwọn Àkọ́lé Àwọn Àmì-ìwé Àwọn Àmì-ìwé Àwọn Ìṣàmúlò-ètò AI Àwọn Àkọ́lé Àwọn Àkọ́lé Díẹ̀ →

Àwọn Àmì-ìwé

Q: Ń lè ṣé ayẹwo àwọn àkọ́lé, àwọn ìròyìn, àwọn àkọ́lé àwọn ìṣàmúlò-ètò nípa ìpelé?

UI ná ìṣẹ̀dà àkọ́lé kan nígbà kan. Fún àwọn ìṣàmúlò-ètò nlà, ìṣàmúlò-ètò JSON+CSV ná gbá ọ̀kan àwọn API láti inú àkọ́lé Python tàbí Node tí n gbọ́dọ̀ kọ́ọ̀kan àwọn ààtò datá rẹ̀. Àwọn ìkúndùǹ: àwọn àmì-à̀ṣe 20,000 fún ayẹyẹ kan, ayẹyẹ 100/aago fún IP fún àwọn òǹlò aimọye, 1,000/aago fún àwọn

Ìlò àìṣé àìṣé Àwọn ìṣàmúlò-ètò 579 Kò ní àmì-ìwé Kò ní àwọn ìṣàmúlò-ètò

Àkọlé àtí a óò ṣàfihàn

0 àwọn àmì-ìwé · 0 Àwọn Àmì-àṣírí · 0 Àwọn àkọlé

Móòdù ìṣàfarawégbè

Àwọn ààyè-iṣẹ́ ìṣàfarawé Àwọn ààyè-iṣẹ́ àwọn ìrànwọ́ Àwọn àwọn ìṣàmúlò-ètò Àwọn àkọlé Àwọn Àkọ́lé

Single positive/neutral/negative verdict with confidence. Best for quick customer-review or NPS tagging.

Ìgbà ìsàlẹ̀-ilà: 2/4

Ààyè-iṣẹ́ Àwọn ìpéwọ̀n Àwọn àgbègbè Forensic

Àwọn ààyè-iṣẹ́

~100 tokens per use

Ṣàfihàn àwọn ìrànwọ́ àyọkà láti inú AI ọ̀fẹ́. Wá àwọn ìṣàmúlò-ètò ìṣàfarawé, ìṣàfarawé, àti àwọn ìṣàmúlò-ètò àìdá.

Bii o ṣe le lo Àwọn Àmì-ìwé

Fi àgbèwọlé rẹ̀ sínú

Ṣàfihàn àkọlé, láti fi fáìlì pamọ́, tàbí láti ṣàfihàn ohun tí o fẹ́. Àwọn iroyin kò ní àwọn ìṣàmúlò-ètò.

Tẹ̀ láti ṣẹ̀dà

AI wa n ṣe iṣẹ́ nínú àwọn ìṣàmúlò-ètò rẹ̀ nínú àwọn ìṣàmúlò-ètò tí o dara jù lọ.

Àwọn Àkọlé

Ṣàfikún, kọ́, tàbí fi pamọ́ sí àwọn ààtò rẹ. Àlàyé fún ìlò inú ẹni-kẹta àti fún iṣẹ́ ọ̀fẹ́.

Lo àtòjọ-ẹ̀yàn yìí láti inú API

Fi àwọn àtòjọ-ẹ̀yàn yìí pamọ́ láti inú ìṣàmúlò-ètò rẹ̀. Àwọn ààyè-iṣẹ́ REST tí a bá fẹ́, àwọn ìṣàmúlò-ètò toké, àwọn SDK àti àwọn ààyè-iṣẹ́ àwọn toké. Àwọn àwọn àwọn owó toké ní àwọn ìṣàmúlò-ètò wẹ́ẹ̀bù.

Àwọn Àkọsílẹ̀ Gba àwọn àyọkà API

curl -X POST https://api.free.ai/v1/chat/ \
  -H "Authorization: Bearer sk-free-..." \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen7b", "messages": [{"role": "user", "content": "Summarize this: ..."}]}'

Àwọn Àṣẹ AI Fúnràn

Àwọn Àtòjọ-ẹ̀yàn Àkọ́lé

Àwọn Àkọ́lé

Àwọn Àmì-ìwé Àwọn Àmì-ìwé

Àwọn Ìṣàmúlò-ètò AI

Àwọn Àkọ́lé

Àwọn Àmì-ìwé - FAQ

Five things, depending on the mode you pick: (1) overall polarity - a -100 to +100 score with confidence; (2) 6-dimensional emotions - joy, anger, fear, sadness, surprise, disgust; (3) aspect-based - extracts named features or entities and scores each separately; (4) sarcasm / subtext - flips surface polarity on detected irony; (5) per-sentence - tags every sentence individually. Pick the mode that matches your use case.

Ya - a 200-word-typical review runs the default Qwen 3 30B model at ~200 tokens, comfortably inside the 6,000 anonymous or 30,000 signed-up daily pool. Forensic-depth aspect analysis on longer texts costs ~600 tokens. No sign-up required to try a few.

O lè ṣé ìṣàwòlẹ̀ nípa ìgbésẹ́ gbogbó (~85% ìgbésẹ́ pẹlú àwọń aláṣẹ́lú eniyan nípa àkọ́kọ́ tí a tí sọ́kalẹ̀), nínú Google nípa ìdájọ́ ìdájọ́, nípa ìdájọ́ tí o jú lọ nípa ìgbésẹ́, nitorí pé a fi àwọn àwọn àkọ́kọ́ ìgbésẹ́-ìdajọ́ pẹ̀lú́ àwọn módèlè̀. Àwọn amọjá nípa ìdájọ́ $1-4 fun gbogbo 1,000 unití. Agbara wa: ìfẹ́rádà ìfẹ́ + ìdájọ́ ìdájọ́-ìpàdé nínú ipe kan, ọ̀fẹ̀, kò sí bọ́tìnì API láti fi hàn.

Yes - switch to the "Sarcasm + subtext" mode. The system prompt explicitly instructs the model to flip surface polarity when it detects irony ("Oh great, another delay" scores negative despite positive-word presence). Accuracy on obvious sarcasm is ~80%; subtle passive-aggression is harder for any model. The "Mixed" verdict + low-confidence score is your flag to read the text yourself.

Instead of one verdict for the whole text, you get separate scores for each entity or feature mentioned. Example: "The hotel was beautiful but the breakfast was cold and the wifi was terrible" yields {hotel: +80, breakfast: -50, wifi: -70}. Ideal for product reviews, restaurant feedback, and NPS comments where customers mix praise and complaints in one paragraph.

99 àwọn ìtàn. Àwọn ìṣàmúlò-ètò ìrànwọ́ àwọn ìrànwọ́ mẹ́tà náà tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n tí wọ́n

Ya - eyi ni ọkan ninu awọn ọran lilo ti o ga julọ. Rọ́ọ̀nù awọn tiketi ti o gba lọ́wọ́lọ́wọ́ nípa "Tiketi Ìdáràn" móòdù dómén; tiketi ti o bá jẹ́ ti òkè -50 nínú ìdáràn nínú ìgbà tí ò bá lọ́wọ́lọ́wọ́, tiketi ti ò bá jẹ́ ti ò bá lọ́wọ́lọ́wọ́ lọ́wọ́lọ́wọ́. JSON tí a bá fi bọ́tìnì ìṣàmúlò-ètò pamọ́ lọ́wọ́lọ́wọ́.

Ya - POST si /v1/gbákọ́/ nípa àgbègbè àtòjọ-ẹ̀yàn àtòjọ-ẹ̀yàn tí ojú-ìwé yìí tí kọ́ (wó àwọn ìṣàmúlò-ètò fún àgbègbè kan). Ǹpa JSON tí a tí kọ́. O dara fún àwọn ààyè tí wọ́ inú àwọn àgbègbè àwọn ọ̀dọ̀ ọ̀dọ̀. Àwọn ìṣàmúlò-ètò àwọn àtílẹ̀, àwọn ìpamọ́ oṣù. Àwọn àkọ́lé nínú /api/.

Confidence reflects the model certainty, not your certainty. Short inputs (<20 words), ambiguous text, and mixed-polarity content score low. A low-confidence positive is not a confident negative - it is "I do not know". Treat any score below 50 as needing human review.

UI ná ìṣẹ̀dà àkọ́lé kan nígbà kan. Fún àwọn ìṣàmúlò-ètò nlà, ìṣàmúlò-ètò JSON+CSV ná gbá ọ̀kan àwọn API láti inú àkọ́lé Python tàbí Node tí n gbọ́dọ̀ kọ́ọ̀kan àwọn ààtò datá rẹ̀. Àwọn ìkúndùǹ: àwọn àmì-à̀ṣe 20,000 fún ayẹyẹ kan, ayẹyẹ 100/aago fún IP fún àwọn òǹlò aimọye, 1,000/aago fún àwọn

MonkeyLearn ti a gba ati ki o ti da duro. Lexalytics ati Repustate gba $500-5,000/mo fun ti o tọ aspekti-based + emotion detection. wa ohun elo ni ọfẹ fun diẹ ninu awọn iṣẹ-ṣiṣe; premium models (GPT-5, Claude Sonnet) wa ni per-call fun awọn ti o ga-stakes lilo awọn ọran bi awọn owo-ọna transcripts.

Àkọlé ní a fi pamọ́ sínú ìrànwọ́ nípa GPU wa, àti àwọn òǹlò àìdára. Àwọn òǹlò tí a tí ìṣàfilọ́lẹ̀ nínú àwọn àkọlé nínú àwọn ààyè-iṣẹ́ wọn ní fun àwọn ọjọ́ 7. A kò gbọ́wọ́lù àwọn àkọlé àwọn àwọn ẹ̀yàn kẹtà láti lò fún ìṣàfilọ́lẹ̀. Fún ìdárànwọ́lù àwọn à

Ṣàfihàn ọ̀fẹ̀: 30,000 tokens/day

Ṣẹ̀dà Àwọn Àkọ́lé

Kò ní àwọn kaadí ẹ̀yàn tí a fẹ́

Bawo ní o ṣe lè fi àwọn ìṣàmúlò-ètò yìí hàn?

Àwọn Àmì-ìwé

Àwọn Àtòjọ-ẹ̀yàn

Bii o ṣe le lo Àwọn Àmì-ìwé

Fi àgbèwọlé rẹ̀ sínú

Tẹ̀ láti ṣẹ̀dà

Àwọn Àkọlé

Lo àtòjọ-ẹ̀yàn yìí láti inú API

Àwọn Àṣẹ AI Fúnràn

Àwọn Àmì-ìwé - FAQ

Kini o n ṣe nínú àwọn ààyè-iṣẹ́ ààyè-iṣẹ́ yìí?

Ó jẹ́ ọ̀fẹ́?

Bawo ni o ṣe tọka si Google Natural Language API tabi AWS Comprehend?

Njẹ o le ṣakoso ifẹnufẹ ati ironu?

Kini àwọn ààyè-iṣẹ́ àwọn ìṣàfarawé àwọn ìṣàmúlò-ètò?

Ìgbà wo nínú àwọn ìrísí-lẹ́tà náà?

Ń lè lò yìí fún ìṣàfarawé àwọn ìṣàmúlò-ètò?

Àwọn ààyè-iṣẹ́ API

Bawo ní àwọn ààtò ìròyìn àìdárà ṣe n ṣiṣẹ́?

Ń lè ṣé ayẹwo àwọn àkọ́lé, àwọn ìròyìn, àwọn àkọ́lé àwọn ìṣàmúlò-ètò nípa ìpelé?

Bawo ni o ṣe le ṣe iyatọ si MonkeyLearn, Lexalytics, ati Repustate?

Àwọn àkọlé mìíràn

30,000 Free Tokens, Gbogbo ọjọ

Wá: 30K tí a fi pamọ́/ọ̀jọ́!

Kò lè fi pamọ́