Interesting APIs 101 x ElevenLabs

Tl;DR

No matter if you are using APIs to BiP, to do some Jira tasks or to bring real life data to your SliDevJS PPT component.

Intro

Whenever you go to a restaurant, you see the menu and order what you want.

You dont care of how its prepared in the kitchen.

Same with API’s: you ask for something, you get it.

And we can do all kind of things with them.

Like using it to push new content/posts programatically to WP or Ghost.

Scrapping APIs

1. Octoparse is a popular no-code web scraping tool designed to automate data extraction from websites. Here are the key facts and features

2. ZenRows is an advanced web scraping SaaS platform focused on making large-scale data extraction from difficult, protected websites easier for developers and businesses.

ZenRows is best for technically adept users needing to scrape protected or dynamic sites robustly and at scale, while not really aimed at true beginners or one-click/no-code users.

3. Exa (https://exa.ai) is a developer-focused Web Search API, AI Search Engine, and Website Crawler designed to power applications needing real-time, accurate, and comprehensive web data. Here’s what makes Exa distinct:

Search built for AI

Summary:

Exa is a powerful API and search engine for developers and AI platforms needing live, reliable access to web data—whether for agentic automation, research workflows, candidate sourcing, market intelligence, marketing analysis, or RAG pipelines.

It’s frequently described as “Perplexity-as-a-service” for customizable, factual, and privacy-conscious web data access.

AI APIs

Avoid LLM lock-in and the recurring subscription bills.

Simply use APIs.

• OpenAI API Keys - https://platform.openai.com/api-keys
• Anthropic - https://console.anthropic.com/settings/keys
• Groq - https://console.groq.com/keys
• For Ollama, you need this setup

Together with Streamlit!

OpenAI

They have released recently o3-mini model.

GPT-4o-mini has been my go to in terms of value per money.

But we have now o1 (mini also) and o3-mini with higher context window (200K)

• https://openai.com/api/
• https://platform.openai.com/docs/models/#current-model-aliases

pip install openai==1.61.0

See the latest openAI models:

import os
from dotenv import load_dotenv
from openai import OpenAI  # pip install openai>0.28

# Load environment variables from the .env file
load_dotenv()

client = OpenAI()
models = client.models.list()
for model in models:
    print(model.id)

https://github.com/JAlcocerT/Streamlit-MultiChat/tree/main/Z_Tests/GrokAPI

DeepSeek via Ollama

Companions for this multichat.

Youtube Summarizer

With Groq API

AIssistant

Leveraging LLMs for several tasks.

Agents with Streamlit

With a very interesting article about agents: https://www.anthropic.com/engineering/built-multi-agent-research-system

• https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools

FULL v0, Cursor, Manus, Same.dev, Lovable, Devin, Replit Agent, Windsurf Agent, VSCode Agent, Dia Browser & Trae AI (And other Open Sourced) System Prompts, Tools & AI Models.

AIssistant

Read the Web!

https://jalcocert.github.io/JAlcocerT/scrap-and-chat-with-the-web/

YT Summarizer

• https://console.groq.com/keys

Ollama API

OpenAI API

We can do many cool things with just simple LLM API calls:

Data APIs

You can get data from other users/services that have collected/processed and expose it for you.

In case you want to create & expose your own API endpoints with code, for example with FastAPI or with Pocketbase BaaS.

Im aware that there are many APIs around: https://github.com/public-apis/public-apis

Even for crypto, with Dune Analytics, official docs

Weather

This was one of the first data I used for a personal project.

Im talking about the python trip planner, which uses the open meteo API to get historical weather data points.

https://api.open-meteo.com/v1/forecast?city=${city}&daily=temperature_2m_max&timezone=auto

Financial Data

This time is going to be yfinance based.

That I also covered recently on this post: https://jalcocert.github.io/JAlcocerT/python-financial-data-with-yfinance/

Formula 1

I got to know recently about: https://www.f1-tempo.com/

Puedes acceder a telemetría de F1 en tiempo real y datos históricos usando APIs y webs especializadas como OpenF1, FastF1 y soluciones comerciales como Formula Live Pulse.

Todas son populares entre fans, analistas y desarrolladores.

Fuente | Gratis | Open Source | De Pago | Notas
———————+———-+—————+———–+———————————————————————————— OpenF1 | Sí | Sí | No | API open source para datos en tiempo real e históricos, usada por desarrolladores. FastF1 (Python pkg) | Sí | Sí | No | Biblioteca Python para acceder a telemetría oficial, análisis y visualización.
Formula Live Pulse | No | No | Sí | Plataforma comercial con API en tiempo real y widgets personalizables.
f1-dash.com | Sí | No | No | Dashboard web gratuito para telemetría en vivo y comparación de sectores.
TracingInsights.com | No | No | Sí | Servicio pago enfocado en análisis y gráficos de telemetría detallados.
F1 TV (App) | No | No | Sí | Servicio oficial de F1 con telemetría, contenido premium y transmisiones.
Ergast API | Sí | Sí | No | API gratuita y open source para estadísticas históricas, no telemetría en vivo.

• OpenF1 API: Ofrece datos en tiempo real y acceso gratuito a históricos, incluyendo telemetría, comunicaciones y vueltas. Ideal para dashboards o análisis detallado. Documentación y ejemplos en openf1.org.
• FastF1 (Python package): Permite extraer telemetría, tiempos por vuelta, posiciones, acelerador, freno y más directamente desde los feeds oficiales. Fácil de usar con Pandas y Matplotlib para visualizaciones. Instalación vía pip o conda.
  • There is also an R version: https://cran.r-project.org/web/packages/f1dataR/index.html
• Formula Live Pulse API: Real-time telemetría, tiempos, análisis de neumáticos y comparaciones entre pilotos. Ofrece widgets personalizables y está disponible para web, iOS y Android.

Opciones en Web y Apps

• f1-dash.com: Dashboard en tiempo real con telemetría y tiempos de carrera online, sector por sector y comparaciones directas.[4]
• TracingInsights.com: Analítica y gráficos interactivos de telemetría y rendimiento, comparando pilotos y vueltas.[15]
• F1 TV App: Oficial de F1, incluye mapas, telemetría, tiempos y está disponible en App Store y Google Play.[19][20]
• DigoRaceDash Lite: App gratuita compatible con F1 24 para mostrar telemetría detallada en tu móvil, sin hardware extra.[8]

Comunidad y Recursos Técnicos

• En Reddit y blogs técnicos, recomiendan FastF1 y OpenF1 como principales fuentes para programadores y analistas. También se menciona Ergast para estadísticas históricas.[6][18]

Estas herramientas te permiten crear dashboards, realizar análisis y estudiar estrategias de piloto en tiempo real con gran profundidad.

2

1. https://github.com/slowlydev/f1-dash

A real-time F1 dashboard

2. https://github.com/br-g/openf1 with official page

As simple as:

curl "https://api.openf1.org/v1/laps?session_key=9161&driver_number=63&lap_number=8"

Social Media APIs

Twitter

Some years back, I was pulling tweets and later applying the detoxify models on it.

• https://docs.x.com/overview

Now the API has changed a bit: https://developer.x.com/en/portal/petition/essential/basic-info

X introduced a tiered pricing model in 2023 when it restructured API access.

• Free tier: Very limited access — typically for small personal projects or testing. Includes minimal monthly post and read limits, and usually no access to enterprise features such as filtered streams or high-volume endpoints.

Get familiar with: https://developer.x.com/en/developer-terms/agreement

If you only need basic functionality (like pulling a few tweets or posting updates), the Free or Basic tier might be enough.

This doesnt mean that you cant use APIfy to pull stuff :)

I intend to use the X API to publish data-driven content focused on the historical performance of companies. 

I want to write better content in less time, helped by Python analytics, so that quality and insightful content can reach further via posts and threads.

Additionally, I will use the API to monitor and understand which types of posts generate the most engagement and resonate best with other users who perform similar analyses.

The goal of this use is to continually improve the quality and relevance of the published content over time, based on observed user interactions and feedback patterns.

I plan to manually replying to comments and discussions on these posts to foster meaningful conversation and community building around such initial data driven posts

But for anything more advanced — analytics, large-scale data collection, or automation — you must pay.

After you submit, you get: https://developer.x.com/en/portal/dashboard

You are allowed to: Web App, Automated App or Bot, Confidential client

1. Pull 100 posts a month
2. Write 500 tweets a month

You will need to create the app:

Sth like: https://grow.libreportfolio.fyi/callback and https://grow.libreportfolio.fyi

LibrePortfolio-X Repository

Source Code - Flask x (Matplotlib Graphs + OpenAI + X API)

DataInMotion Repository

A good combo - Matplotlib Graphs + X API

So that you get a clientID and clientSecret.

Here is your OAuth 2.0 Client ID and Client Secret https://github.com/xdevplatform/Twitter-API-v2-sample-code

Which is not a bearer token https://docs.x.com/fundamentals/authentication/oauth-2-0/bearer-tokens

Time to create some create_post_simple.py and bundle it around a quick Flask app.

#git clone https://github.com/JAlcocerT/DataInMotion
cd OpenAI-Twitter-API
docker compose up -d
#sudo docker compose down

Go to http://192.168.1.2:5033/ pointing to x-api-tweet-creator:5000

#docker network ls | grep cloudflared_tunnel
docker network connect cloudflared_tunnel x-api-tweet-creator #connect
#docker inspect x-api-tweet-creator --format '{{json .NetworkSettings.Networks}}' | jq

networks:
  #adventurelog_net:  # Define the internal network
  # nginx_nginx_default:
  #   external: true #for https
  cloudflared_tunnel:
    external: true # Mark the 'tunnel' network as external

And you will get a twitter post:

{"access_token_redacted":"abcdef...","has_refresh_token":true,"message":"User authenticated","token_type":"bearer"}

So i could not resist to put together everything:

git clone https://github.com/JAlcocerT/LibrePortfolio-X.git

cd LibrePortfolio-X
#uv run app.py

sudo docker compose up -d
#sudo docker compose build --no-cache
#sudo docker compose down

And add x-gen-ui:5088 to grow.libreportfolio.fyi

Then, each generation goes to https://grow.libreportfolio.fyi/preview/06cbbc68a88b and 4 graphs are uploaded with the comment.

Threads

The Threads API by Meta has specific limits designed to control usage and maintain platform integrity.

• Posting Limit: Up to 250 posts per 24-hour period per Threads account via the API.
• Replies Limit: Up to 1,000 replies per 24-hour period per account.
• Rate Limits: The number of API calls (requests) allowed per app or token is controlled by Meta’s Graph API rate limiting, which varies based on your app status and usage history.
• Business Verification: For production-level or larger scale usage, an approved and verified Meta business account is required.
• Data Access Limits: Access to user and content data through the API is governed by privacy and consent requirements, with limitations on the amount and type of data you can pull per request and over time.

These limits ensure fair use and prevent abuse or spamming via automated tools on Threads. For more details, see the official Meta documentation: developers.facebook.com/docs/threads and the Graph API rate limiting page.

If you need more specific technical limits or help navigating them, I can assist further.

Random APIs

If the formula 1 data was not enough:

1. Transform PDFs into markdown locally and quickly is possible thanks to kreuzberg as seen with the Jira post.

Checking APIs

As we saw during the FastAPI post, here, there are few tools to check/test/interact with APIs.

Among all of them, is worth to get to know how to use Hoppscoth.

How to use Hoppscotch

When playing/creating APis…

wget https://github.com/hoppscotch/releases/releases/latest/download/Hoppscotch_linux_x64.deb
sudo apt install ./Hoppscotch_linux_x64.deb #https://hoppscotch.com/download

Other API Testing Tools

1. https://github.com/httpie/desktop

wget -P ~/Applications https://github.com/httpie/desktop/releases/download/v2025.2.0/HTTPie-2025.2.0.AppImage
#httpie

🚀 HTTPie Desktop — cross-platform API testing client for humans. Painlessly test REST, GraphQL, and HTTP APIs.

2. Reqable - Reqable implements the core features of traffic analysis and API testing, and deeply integrates them.

One app is worth multiple apps: Advanced API Debugging Proxy and REST Client

Reqable = Fiddler + Charles + Postman

flatpak install flathub com.reqable.Reqable

3. https://github.com/mountain-loop/yaak/

MIT | The most intuitive desktop API client. Organize and execute REST, GraphQL, WebSockets, Server Sent Events, and gRPC 🦬

Conclusions

The problem today is not to have options, but too many:

• Smithery: An open platform and registry for managing, sharing, and deploying modular agentic tools (MCP servers) that extend AI assistants with real-world actions and data.

See https://smithery.ai/playground

• Octoparse: A no-code/low-code web scraping platform for visually extracting structured data from websites, ideal for non-programmers and small businesses.
• ZenRows: A powerful API-based web scraper built for developers to extract data from dynamic, protected, or anti-bot websites at scale.
• Pipedream: A developer-oriented cloud automation platform for building, integrating, and running event-driven workflows that connect APIs, apps, and custom code.
• Exa: An advanced web search API and site crawler that powers AI and developer workflows with live, customizable, and accurate web data retrieval.

FAQ

I know that some marketers are using…

PipeDream

Pipedream is a cloud-based automation and integration platform designed for developers and technical users. It allows you to create workflows that connect APIs, databases, cloud apps, and event sources with minimal infrastructure setup.

Key points about Pipedream:

• Low-Code/Code-Optional: You can build workflows using a mix of pre-built steps (“actions”) and your own custom code in Node.js, Python, Go, or Bash. It’s not “no-code”—some programming/developer know-how is helpful for advanced automations.[1][2]
• Event-Driven Workflows: Automate processes triggered by events (webhooks, schedule, app activity, HTTP requests, etc.). A flow could start with a new form submission and end with an email sent, data added to Google Sheets, or anything else API-powered.[2][3][4]
• Massive App Integration: Natively connects with thousands of popular SaaS apps and their APIs—think Slack, Notion, AWS, Google, GitHub, Trello, databases, and custom endpoints. Useful for automating business, marketing, DevOps, or productivity tasks.
• Developer-Focused Features: Built-in code editor, version control, real-time debugging, execution history, error handling, and easy management of connections and secrets.
• Use Cases: ETL (Extract, Transform, Load) between systems, automation of business processes, scheduled reporting, CRM updates, social media posting, API orchestration, internal tool building, and more.[5][2]
• Pricing: Has a generous free tier, but limits on workflow invocations and event history; paid plans for heavier usage.[3]
• Comparison: Unlike Zapier or Make.com, Pipedream exposes the underlying logic/code—you get the power of serverless backend but managed hosting and easy debugging. It appeals to users who want flexibility, speed, and API agility.[6][1]

Summary:

Pipedream is ideal for developers wanting rapid, flexible automation between cloud services, APIs, and custom code, without managing servers.

It’s fast for prototyping, robust for event-driven scripting, and widely used for DevOps, SaaS, and workflow automation.

Not a true no-code solution; best suited for technically comfortable users.

Smithery

Smithery (or Smithery.ai) is a developer-focused platform for building, discovering, and managing MCP (Model Context Protocol) servers—modular tools that extend the capabilities of AI assistants, agentic systems, and language models.

Extend your AI with 7572 skills built by the community

• MCP Registry: Smithery acts as a central hub (sometimes called “Google for MCPs”) where developers can discover and share MCP-compliant AI tools—these are small servers or extensions that connect AI models to real-world data, services (like GitHub, Gmail), files, APIs, and much more.[1][2][3][4]

• Agentic AI Tooling: MCP servers let AI agents perform specialized tasks, such as web search, code execution, file management, memory recall, automation with external services, cloud browsing, and so on. Smithery routes each AI agent to the right MCP server for its needs via a dynamic toolbox.

• Local & Hosted Installation: Servers can be run locally (with CLI and Docker support) or hosted in the cloud. Smithery enables easy installation and management—often a single command will set up an MCP server, inject credentials, and register the endpoint for use by AI agents in your environment.[2][5][6]

• Developer Integration: Designed for modern dev ecosystems, Smithery integrates with GitHub (for CI/CD and code sharing), VS Code, Docker, Node.js, and major LLM platforms (Claude, Cursor, etc.). There’s robust support for versioning, automated deployment, and environment variable configuration.

• Privacy & Security: Credentials/tokens (like your GitHub PAT) are generally managed locally and not stored on Smithery’s servers. Audit logs, team permissions, and best practices help secure collaborative workflows.[7][2]

• Accessibility & Scale: The platform is open source (Apache 2.0 license), with frequent updates and a fast-growing community registry—recent counts claim thousands of MCP tools and daily usage in the tens of thousands.[4][1][7]

• Use Cases: Example MCP servers include web search plugins (like DuckDuckGo), automation helpers for software planning, code management, external database connectors, browser automators, and more. Developers can build new MCP servers for virtually any system their agents need to access.[5][6][8]

Summary:

Smithery is the “app store” or registry for agentic AI extensions, enabling powerful automation and integration between LLM-powered agents and real-world apps/services. It’s especially useful for developers creating advanced personal finance, coding, or infrastructure management systems with agentic or autonomous logic.[3][1][2][4]

10 11

Audio APIs

Powerful Translator: https://www.deepl.com/en/translator

If you like Podcasts

• https://www.gitpodcast.com/rybbit-io/rybbit

https://jalcocert.github.io/JAlcocerT/local-ai-audio/

OpenAI TTS and Transcription Project

Speech rAIter Post

NEW Speech Raiter

Streamlit Speech rater - Source Code on Github

11Labs x MCP

ElevenLabs already has its own MCP Server:

But I felt like trying elevenlabs and test how they can do audio was already overdued.

So…

See how crazy it sounds:

Say thanks to this vibecoded HUGO shortcode for audio btw

Sources from ./static/audio/

Python CLI Tools

• https://pypi.org/project/questionary/ - Create CLI utilities with python
• https://docs.pytest.org/en/stable/

FAQ

Other Tools

• https://github.com/Chivo-Systems/Shinar/

gpl 3.0 | Open Source AI Calling Transcriptions, Summaries, and Analytics built on OpenAI Whisper

• https://github.com/getmaxun/maxun

agpl | No Code Web Data Extraction Platform • Turn Websites To APIs & Spreadsheets In Minutes

• https://github.com/clusterzx/paperless-ai

MIT | An automated document analyzer for Paperless-ngx using OpenAI API, Ollama, Deepseek-r1, Azure and all OpenAI API compatible Services to automatically analyze and tag your documents.

• https://github.com/esxr/langgraph-mcp

MIT | LangGraph solution template for MCP

Creating TwitterOpenAIMatplotlib Bot

git init
git branch -m main
git config user.name
git config --global user.name "JAlcocerT"
git config --global user.name
git add .
git commit -m "Initial commit: Starting LibrePortfolio x Twitter OpenAI Matplotlib Bot"

#sudo apt install gh
gh auth login
gh repo create LibrePortfolio-X --private --source=. --remote=origin --push