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

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.

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. See PyGPT

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/: 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
Deeper Hermes via LMStudio ( https://lmstudio.ai/ ): LM Studio is for using some type of models locally similar to ollama. Deeper Hermes has automatic switch between direct answer and deep thinking based on context which makes it really great.
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

Anonymous