Overall list of useful AI tools

Home > Web based AI tools > Overall list of useful AI tools

Overall list of useful AI tools

perplexity.ai - Perplexity

For very accurate answers especially pro version. It can also go online and look at latest information for answering. Integrate via ThunderAI to Thunderbird using Thunder AI Thunderbird plugin based access to perplexity API

Comet browser from perplexity available only for Windows and Mac and for now is available based on invitation. In windows during usage it was able to do some tasks eg go through emails. But when asked to download images from Whatsapp already open in browser tab it was not able to do so.

chat.openai.com - Chat GPT

Good for writing or creative tasks. Integrate via Aify to Thunderbird. See Aify Thunderbird plugin based access to OpenAI API It is multimodal so that it can generate images, work with audio and video. It can understand spreadsheets, etc. It can be used including image generation and recognition via API on local system using PyGPT

gemini.google.com ( https://aistudio.google.com/ ) - Gemini 2

It can be used to learn tool by sharing screen. We can also do live interaction with it completely via voice.

eraser.io/diagramgpt

This is good at drawing diagrams

Ollama

This can run locally on system as explained at Ollama. We can integrate this with Thunderbird via ThunderAI plugin as it supports local OpenAI compatible API. See Using Ollama with Thunderbird ThunderAI plugin

usebubbles.com

AI notetake and recording tool for meetings. No need of local agent and works completely via browser even in Linux.

notebooklm.google.com

We can upload sources including pdf, word, youtube links, etc. and then get replies to queries only based on uploaded sources. This can build a podcast or executive summary based on uploaded documents. We can also get timeline of events, if applicable.

PyGPT

Do ChatGPT API based queries and tasks locally on system. This has ability to run local commands. There are also options for controlling mouse and keyboard, accessing Internet, generating images, voice based input/output, file I/O, Vision (Image recognition), etc. It tries to control screen via keyboard and mouse but that is not that great.

make.com

Automated workflows to do some task based on some condition. I have tried multiple others for some email automation but this is the only one that worked smoothly.

• I configured it to check emails via IMAP, send the body of email to perplexity along with custom message to draft reply to email, then send the perplexity output directly to sender via SMTP.

missiveapp.com

Very good tool for checking email and it has equally good mobile app for Android. It is good for shared team inboxes. There is private chat per email. There is option to compose with OpenAI integration. It has very good integration with lot of other tools.

Fireflies.AI (https://fireflies.ai/apps)

AI tool for automatically joining meetings, taking notes, and providing transcripts and summaries. This has unlimited audio recording for recording in-person meetings via phone app. By default there are only 8 recordings or 800 minutes. However, if we enable join all / share all settings then we get unlimited recording. This is fine if we dont use that account anyway.

https://notion.so

For notes and tasks planning. It is like having emacs org mode in web browser that can be shared with others, linked to other documents and web pages, include images, etc.

https://airtable.com

For organization information in large tables format with automation / AI and hooks. These hooks can even be called from make.com. The mobile app of airtable is also great and allows editing one row at a time with very easy to use UI. We can create views and share only subset of rows as a link with others.

Otter

Decent recording capabilities but free version of fireflies has more features.

chat.deepseek.com ( esp. DeepSeek-v3-R1 Model)

Very good reasonsing model available even with free account

https://excalidraw.com/

Very good tool to draw diagrams especially using text to generate image. Best part is there is no registration or login. Just open page and use it. Also there is option to save drawing on system or share via link or export. Further we can have a live collaboration session also.

Open Interpreter

Using this we can type what we want and it will build code and commands on fly, ask our permission and then execute the same to get desired result. The advantage of this is that this works with Ollama models also. See Open Interpreter

Self Operating Computer

We can use it to completely control computer using instructions. For example it can open browser and search. See Self Operating Computer

sgpt

This can be used to use OpenAI API directly from shell. We can do quick search and get reply. We can give image as input and write query for the image. We can also search for shell commands to be executed and it will generate command and take permission before running it (Similar to interpreter). See sgpt

https://pinokio.computer/ Pinokio

AI browser to look at various AI models to run them locally on computer. We can use this to install Open Web UI

Open Web UI ( https://github.com/open-webui/open-webui )

This is similar to PyGTK but has option to use local LLM which we are unable to do via PyGTK. After installation it was able to detect local ollama server and all installed models automatically. The advantage is ease of use same as CLI use of ollama models but also has history same as PyGTK without being very heavy. It also has OpenAI integration via key so that both OpenAI and ollama models can be run side by side in same tool. Further working with vision models is easier as we can upload images via UI. Might work properly in Chrome compared to firefox for UI features

Browser use ( https://github.com/browser-use/browser-use )

Does not work on Rocky 9.x but might work on other OS

Forge

This is allowing creation of images with just 3GB of VRAM using flux.

Hugging face ( https://huggingface.co/ )

Platform for multiple AI tools / models. We can see models and download them to run locally. Or we can use inference engines from replicate.com etc. to run the models with high GPU VRAM, system RAM, etc. A few models can be tried directly on site for free.

Replicate ( https://replicate.com/ )

Platform to run various AI models by renting machines per second including GPU. There are readymade models to run on top of their hardware. I have not tried paid account of this yet.

Fal ( https://fal.ai/ )

Similar to replicate.com but mostly for media (Images, Video etc.) and not open for running other models eg Deepseek-r1. I have not tried paid account of this yet.

Grok ( https://grok.com/ )

Very good AI with deep reasoning and think options while also being free to use for foreseeable future.

N8n ( https://n8n.io/ )

We can automate workflows locally on system similar to using ollama locally without using make.com or other portals. See N8n

Perplexica ( https://github.com/ItzCrazyKns/Perplexica )

This is local installation simliar to WebUI but it integrates RAG with help of searxng based search. The answers are better than Msty etc. other products that use same models (Eg deepseek-r1:1b or gemma3:1b) but they are not as accurate as public perplexity.ai site. In terms ease of installation and UI features is Msty much better then this. See Perplexica

Msty ( https://msty.app/ )

This is local app, perhaps better than OpenWebUI, Perplexica, PyGPT, etc. in terms of starting faster and supporting large no. of AI based on API (including perplexity, remote ollama, OpenAI, etc.). We can resend the query and look for other results. This is supposed to have RAG with Internet search + spaces where we can upload documents. We can edit the generated response if we want the follow up question to consider edited response. See Msty

Some how quality of answer and accuracy of perplexica while using same models eg deepseek-r1

1b or gemma3:1b with RAG / Internet search is much better than Msty.

OpenRouter ( https://openrouter.ai/ )

We can use OpenRouter for API using tools such as Msty. With help of OpenRouter no need to pay so many API providers separately. We can use perplexity, OpenAI, x.ai, etc. including new ones such as Kimi K2 directly using openrouter using a single tool (eg Msty) and using a single recharge.

Google Veo

Generates only 7-8s clip in India without any audio via Web UI using Veo 2. Unless there is global release for Veo 3 this is not useful for India region at least.

videogen.ai ( http://videogen.ai/ ) - Very useful to generate videos with small stock clips at back, text to speech audio with captions.

You can consider recording screen and audio output via OBS Studio to get the generated clip captured properly. In OBS studio get audio captured post noise cancellation from output. Also get screen captured with crop settings to ensure that only video part of the play is recorded. We can setup shortcuts for start/stop/pause/unpause recording and then go to videogen.ai. Start recording just before playing and then stop as soon as clip ends.

Open Code Terminal ( https://opencode.ai/ ) - For vibe coding using Grok-4 and perhaps might support KimiK2 also. But the overall quality was not great and the API dollar usage was quite high.

Gemini CLI - This is good for Vibe coding and also free. But it uses Gemini only. We cannot use any other AI for backend.

Qwen CLI - This is also for vibe coding and it can work with local ollama models such as qwen3. Again output is not good compared to Gemini CLI.

Home > Web based AI tools > Overall list of useful AI tools