bloopcommitsNaya AI model try karna ho aur pehle subscription ka page khul jaye, to mazaa kirkira ho jaata hai. Google AI Studio ek aisi jagah hai jahan Gemini models ke saath kaam shuru hota hai. Is guide mein hum dekhenge ki official pages ke hisaab se kya free hai, kya nahi, andar jaane ke steps kya hain, teen features jo docs mein likhe hain (system instructions, temperature, context window), ek test plan, aur sabse zaroori cheez: free tier ka data kaise use hota hai.
Google AI Studio kya hai
Google ke apne landing page par likha hai ki Google AI Studio Gemini ke saath building shuru karne ki jagah hai (Gemini = Google ke multimodal AI models ka family). Landing page par "Sign in to Google AI Studio" ka button hai, yaani andar jaane ke liye Google account se sign in karna padta hai. Studio ka asli workspace login ke peeche hai, isliye Reel mein sirf public page dikhaya gaya tha, aur is guide mein bhi koi screenshot ya copy nahi hai.
Free kya hai, kya nahi
Gemini API ke pricing page par likha hai: "Google AI Studio usage is free of charge in all available regions." Wahi page Free aur Paid tier ko alag dikhata hai. Neeche ki table usi page aur billing page ki baaton ka seedha saransh hai.
| Free tier | Paid tier | |
|---|---|---|
| Kharcha | Free of charge (input aur output tokens bhi) | Billing account link karna padta hai |
| Models | Sirf kuch models tak access | Zyada aur advanced models tak access |
| Limits | Free tier ke rate limits | Higher rate limits |
| Tumhara content products improve karne mein use hota hai? | Haan (pricing page ki row: Yes) | Nahi (pricing page ki row: No) |
- Rate limits ka matlab: ek time-frame mein tum kitne requests bhej sakte ho. Ye model ke hisaab se alag hoti hain.
- Docs khud kehte hain: limits guaranteed nahi hain aur actual capacity badal sakti hai.
- Isliye koi number yaad mat rakho. Hamesha official rate limits page dekho.
Setup: step by step
- Browser mein Google AI Studio ka official landing page kholo (link neeche "Official links" mein hai). Ye page public hai.
- "Sign in to Google AI Studio" button dabao aur apne Google account se sign in karo.
- Andar ka layout Google badalta rehta hai, isliye yahan koi screenshot nahi hai. Docs jin cheezon ke naam lete hain, unhe dhoondho: prompt, system instructions, temperature.
- Pehle sab settings default par chhodo aur ek chhota public text lekar prompt chalao.
- Phir ek baar mein sirf ek setting badlo (neeche test plan hai) aur dekho jawab mein kya farak aaya.
Teen features jo Gemini docs mein likhe hain
1. System instructions
Docs ke hisaab se: "You can guide the behavior of Gemini models with system instructions." Matlab prompt se pehle tum model ko ek standing instruction de sakte ho, jaise uska role. (Docs ka apna example bhi ek role hai.)
2. Temperature
Docs ke hisaab se: "The temperature controls the degree of randomness in token selection." Kam temperature un prompts ke liye achha hai jinhe zyada deterministic ya kam open-ended jawab chahiye, aur zyada temperature se zyada diverse ya creative results aa sakte hain.
3. Long context window
Docs ke hisaab se kai Gemini models ke context window bade hote hain, aur long context window se naye use cases khulte hain. Seedhi bhaasha mein: lamba input ek saath diya ja sakta hai. Kisi model ki exact limit jaanni ho to docs ka Models page dekho; wo limit badal sakti hai.
- Docs kuch naye Gemini models ke liye salah dete hain ki temperature jaisi sampling settings default par hi rakho.
- Docs ke hisaab se temperature ko default se neeche karne se kabhi kabhi unexpected behavior (jaise looping ya kharab performance) ho sakta hai.
- Isliye pehle default se shuru karo, aur settings ek ek karke badlo.
Ek banaya hua example (asli output nahi)
Neeche ka example maine banaya hai, ye kisi product ka asli screen ya output nahi hai. Idea ye hai ki system instruction role deta hai, aur prompt asli sawaal.
System instruction: Tum ek teacher ho. Simple Hindi mein samjhao. Prompt: Photosynthesis kya hota hai? Ek chhota example do.
Test plan: ek baar mein ek setting
| Run | Kya badlo | Kya dekho |
|---|---|---|
| 1 | Kuch nahi, sab default | Jawab ka tone aur length. Ye tumhari baseline hai. |
| 2 | Sirf system instruction jodo | Kya jawab ab us role mein aaya? |
| 3 | Sirf temperature thoda kam karo, wahi prompt 3 baar chalao | Kya jawab ek jaise aa rahe hain? |
| 4 | Sirf temperature thoda zyada karo, wahi prompt 3 baar chalao | Kya jawab alag alag aa rahe hain? |
Ye sirf ek test plan hai. Docs temperature ko randomness ka level batate hain, lekin tumhare prompt par kitna farak padega, ye tumhe khud dekhna hoga.
Catch: free tier mein tumhara content
Gemini API ke pricing page ki row "Used to improve our products" mein Free tier ke saamne "Yes" likha hai aur Paid tier ke saamne "No". Terms page ye aur detail mein kehta hai.
- Jo services free of charge hain (jaise direct Google AI Studio use aur Gemini API ka unpaid quota) wo Unpaid Services hain.
- In mein Google tumhara submit kiya content aur generated responses ko Google ke products aur machine learning technologies ko provide, improve aur develop karne ke liye use karta hai.
- Quality ke liye human reviewers tumhara API input aur output padh sakte hain. Terms ke hisaab se wo data tumhare Google Account, API key aur Cloud project se alag karke dikhaya jaata hai.
- Terms ki seedhi line: "Do not submit sensitive, confidential, or personal information to the Unpaid Services."
Terms page mein ye bhi likha hai ki European Economic Area, Switzerland ya United Kingdom mein Paid Services wale data terms sab services par lagte hain, Google AI Studio samet. Aur billing page ke hisaab se, billing wale API project hone par AI Studio prompts bhi Paid Services terms ke andar aate hain. Tumhare case mein kya lagta hai, ye terms page aur billing page se khud confirm karo.
Aam galtiyan
- Private, client ya company ka data free tier mein daal dena. Terms page isse mana karta hai.
- "Free" ko "sab kuch, hamesha, bina limit" samajh lena. Free tier mein kuch hi models milte hain aur limits hoti hain.
- Limits ka koi number yaad rakh lena. Docs kehte hain limits guaranteed nahi hain aur badal sakti hain.
- Ek saath kai settings badal dena, phir samajh na paana ki farak kisse aaya.
- Ye maan lena ki sign in ki zaroorat nahi. Google account se sign in karna padta hai.
Checklist
- Official landing page se hi sign in kiya.
- Pehle sirf public ya nakli text se test kiya.
- Settings default par chhodkar baseline jawab dekha.
- Ek baar mein ek hi setting badli.
- Pricing page ki "Used to improve our products" row padh li.
- Rate limits ka current page dekha, purana number yaad nahi rakha.
- Google AI Studio ka use free of charge hai, par free tier mein kuch hi models, limits aur paid tier ka option hai.
- System instructions = model ka role. Temperature = randomness ka level. Context window = lamba input ek saath.
- Free tier ka content Google ke products improve karne mein use ho sakta hai: private, client ya company data mat daalo.