Tool Guides & Tutorials

AI Chat is a fully on-device conversational AI. No internet required once the model is downloaded. Ask questions, brainstorm, outline projects, or explore ideas freely.

How to use it

1. Open AI Chat from the tool directory or the homepage.
2. Wait for the model to download (first use only — usually 50–300 MB). A progress bar shows the download status.
3. Type your message in the chat box and press Enter or click Send.
4. Continue the conversation naturally. The model remembers the current conversation context.
5. Start a new conversation with the + button, or rename/delete conversations from the sidebar.

Practical examples

• "Explain quantum computing in simple terms for a 12-year-old."
• "Give me 10 blog post ideas about sustainable living."
• "I'm writing a cover letter for a marketing role. What should I emphasise?"

Limitations

• On-device models are smaller than cloud models (e.g. GPT-4). Complex reasoning tasks may be less reliable.
• No internet access — cannot look up current events, prices, or live data.
• Conversation memory is limited to the current session. Clearing browser storage removes saved chat history.

Turn a topic or brief into a first draft. Useful for blog posts, articles, product pages, reports, or any long-form writing project.

How to use it

1. Enter your topic or brief in the input field. Be specific about the audience, purpose, and length you want.
2. Click Run. The model will generate a draft based on your input.
3. Copy the output and refine it in your own editor. Treat the output as a first draft, not a final product.

Example input

Limitations

• Outputs are generic starting points and will need personalisation and fact-checking.
• The tool does not have access to up-to-date information or sources.

Correct spelling, grammar, and punctuation in a short passage without sending your text to any external service.

How to use it

1. Paste or type your text (up to a few paragraphs) into the input box.
2. Click Fix grammar.
3. Review the corrected version. Changes are highlighted where possible.

Get a shorter version of an article, set of notes, or passage. Ideal for research, studying, or quickly digesting long reports.

How to use it

1. Paste the text you want to summarise into the input field.
2. Click Summarise. The model condenses the main ideas.
3. If you want a shorter or more detailed summary, mention the desired length in a follow-up or trim/expand manually.

Sample result

Limitations

• Very long texts may need to be split into chunks.
• Summaries may occasionally omit nuanced details. Always read summaries alongside the original for important decisions.

Determine whether English text sounds positive, negative, or neutral — useful for customer feedback, reviews, or social media monitoring.

How to use it

1. Paste your text (a sentence to a few paragraphs).
2. Click Analyse. The tool returns a sentiment label (Positive / Negative / Neutral) and a confidence score.

Extract written text from an image — photos of documents, screenshots, scanned pages, signs, or handwritten notes.

How to use it

1. Click Upload image and select your file (JPEG, PNG, WEBP, etc.).
2. Click Extract text. Tesseract.js runs entirely in your browser.
3. Copy the extracted text from the output panel.

Tips for best results

• Use high-resolution images (at least 150 DPI for scanned documents).
• Ensure text is horizontal and well-lit. Skewed or low-contrast images reduce accuracy.
• Works best with printed, clearly typed text. Handwriting accuracy varies significantly.

Limitations

• Complex layouts (multi-column PDFs, tables) may not be perfectly preserved.
• Accuracy decreases with poor image quality, unusual fonts, or heavily stylised text.
• Your image is never uploaded to any server — processing is fully local.

Remove the background from a portrait photo to create a transparent PNG. Useful for profile pictures, team pages, product shots, and presentations.

How to use it

1. Upload a portrait photo where the subject is clearly separated from the background.
2. Click Remove background. Processing runs locally (may take 5–15 seconds).
3. Download the resulting transparent PNG.

Limitations

• Works best with clear portraits where the subject is well-defined against a contrasting background.
• Complex backgrounds (e.g. detailed foliage, similar-coloured clothes) may produce imperfect cutouts.
• Results are best reviewed and may need minor touch-up in an image editor for professional use.

Convert a spoken audio recording to text — meetings, voice notes, lectures, or interviews. Everything processes in your browser.

How to use it

1. Click Upload audio and select an MP3, WAV, or OGG file (under ~5 minutes for best performance).
2. Click Transcribe. The Whisper model runs locally — first use downloads the model file.
3. Review and copy the transcript from the output panel.

Tips for best results

• Use clean recordings with minimal background noise.
• Speak clearly and at a moderate pace.
• For long recordings, split them into 3–5 minute segments for faster and more accurate results.

Limitations

• Strong accents, technical jargon, or multiple overlapping speakers reduce accuracy.
• The on-device Whisper model used is the "tiny" or "base" variant — accurate for general English, but not as precise as the larger cloud versions.
• Processing time depends on your device speed. Older computers may take longer.

Draft a professional cover letter from your experience and the job details you provide.

How to use it

1. Enter the job title and company name you are applying for.
2. Briefly describe your relevant experience and key skills (a few bullet points is enough).
3. Click Draft letter. Review, personalise, and adapt the output before sending.

Example input

Turn your notes or a source passage into Q&A flashcards for quick revision.

How to use it

1. Paste your notes, textbook excerpt, or article into the input field.
2. Click Make flashcards.
3. The tool generates a set of question-and-answer pairs. Copy them into your preferred flashcard app (Anki, Quizlet, etc.) or print them out.

Sample output

Chat with Google's Gemini model using your own free API key. Gemini is a powerful large language model suited for complex reasoning, creative writing, and multi-step tasks.

Getting a free API key

1. Visit Google AI Studio and sign in with your Google account.
2. Click Create API key. Copy the key.
3. Return to Free AI Harbor, open Gemini Assistant, paste your key, and click Save key.
4. Your key is stored only in this browser tab's memory and cleared on page reload.

Limitations

• Requires an internet connection.
• Free tier has rate limits (requests per minute and per day). Heavy use may hit limits.
• Your inputs are sent to Google's servers. Do not submit sensitive or confidential personal data.