AI tools
Google Gemini CLI
An open-source terminal coding agent. Its current README directs unpaid and Google One users to Antigravity CLI; verify current access before choosing it.
Ollama model runtime
A runtime for working with open models through a command line or API, with local and cloud options.
Open WebUI workspace
A self-hosted chat workspace with documented support for local models, document retrieval and compatible model APIs.
Langflow visual builder
A visual environment for composing AI workflows and exposing them through APIs or MCP tools.
Dify AI application platform
A platform combining visual AI workflows, document retrieval, model connections and application APIs.
Hugging Face Transformers
A Python library with model pipelines for text, audio, vision and multimodal tasks.
Microsoft MarkItDown
A Python utility that converts supported documents into Markdown for text analysis and AI workflows.
Use cases
Explore your documents with source-aware chat
Open WebUI documents retrieval over uploaded documents, allowing you to ask questions alongside your own materials.
Turn documents into readable Markdown
We converted a sample Word document with Microsoft MarkItDown. Headings, lists, tables, dates and links survived; the page header and footer did not. Here is how to check your own result.
Prototype an assistant for a document collection
Dify documents a retrieval pipeline that spans document ingestion and retrieval within an AI application.
Explore speech-to-text with an open model pipeline
The Transformers README includes a speech-recognition pipeline example using a Whisper model.
Try a model conversation on your own infrastructure
Ollama documents a chat API and command-line model workflow, with local execution available for compatible models.
Comparisons
Codex vs Gemini CLI: access, tools and constraints
A research-based comparison of coding-agent entry points, documented tools and account requirements.
Dify vs Flowise: two visual AI building approaches
Compare documented product scope and deployment considerations without assuming that visual building removes operational work.
Ollama and Open WebUI: runtime or workspace?
These tools occupy different layers: one runs models, while the other provides a workspace that can connect to a model backend.
Langflow vs Dify: composing an AI application
Two documented approaches to visual AI workflows, compared around integration and retrieval needs.
AI News
Dify 1.17.1 documents narrower knowledge-base API access
The release describes dataset-scoped keys, workflow keyboard controls and an important self-hosted upgrade caveat.