API plumbing
Field notes on coding agents, model choice, token math and API plumbing.
-
terminal-health-checks-and-safe-model-fallbacks-102.mdTerminal Health Checks and Safe Model Fallbacks
Build bounded endpoint probes, classify failures, and test a two-model fallback from Bash without hiding authentication or quota problems.
-
terminal-proof-api-retries-in-bullmq-workers-100.mdTerminal-Proof API Retries in BullMQ Workers
Build a BullMQ worker with stable job IDs, durable results, bounded retries, and a deliberate crash test for duplicate-safe processing.
-
switch-api-providers-when-you-change-project-folders-98.mdSwitch API Providers When You Change Project Folders
Set up project-local environment files, select model IDs, and switch API providers without changing your Python application code.
-
corporate-proxy-setup-for-python-sdks-in-your-terminal-97.mdCorporate Proxy Setup for Python SDKs in Your Terminal
Configure proxy variables, corporate CA trust, and an OpenAI SDK request through Fluxo without disabling TLS verification.
-
terminal-request-logs-that-keep-llm-prompts-private-96.mdTerminal Request Logs That Keep LLM Prompts Private
Build an allowlisted LLM request logger with exact terminal commands, safe error handling, private files, and a quick leak check.
-
from-chat-response-to-valid-json-in-your-terminal-94.mdFrom Chat Response to Valid JSON in Your Terminal
Send a JSON Schema to a chat endpoint, inspect the response, and validate the result locally with curl, jq, and Python.
-
api-keys-from-your-terminal-to-ci-and-production-90.mdAPI Keys from Your Terminal to CI and Production
A practical walkthrough for keeping API keys out of Git, shell history, CI logs and production configuration, with commands you can copy.
-
inspecting-streaming-usage-tokens-in-your-terminal-92.mdInspecting Streaming Usage Tokens in Your Terminal
Capture an SSE response, parse its final usage event, and preserve cache token details with curl and a small Python script.
-
a-tiny-ai-cli-for-your-terminal-and-editor-91.mdA Tiny AI CLI for Your Terminal and Editor
Build a Python CLI for an OpenAI-compatible API, discover model IDs, pipe files into prompts, and keep credentials out of source control.
-
long-streams-without-blind-retries-in-your-terminal-89.mdLong Streams Without Blind Retries in Your Terminal
Set explicit streaming timeouts, disable hidden retries, and preserve partial output with a runnable Python client for Fluxo.
-
terminal-checks-for-model-endpoint-latency-and-errors-85.mdTerminal Checks for Model Endpoint Latency and Errors
Build a repeatable terminal probe to track endpoint latency, HTTP failures, and changes in model performance.
-
claude-messages-from-your-terminal-with-curl-88.mdClaude Messages from Your Terminal with curl
Call the Anthropic Messages API through Fluxo with curl, inspect usage, send conversation history, and stream responses without an SDK.
-
call-generatecontent-from-your-terminal-with-fluxo-87.mdCall generateContent from Your Terminal with Fluxo
Send native generateContent requests through Fluxo with curl and Python, inspect response text, and debug common request failures.
-
migrating-a-terminal-script-to-the-responses-api-86.mdMigrating a Terminal Script to the Responses API
Move an OpenAI SDK script from Chat Completions to Responses through Fluxo, with runnable commands and migration checks.
-
a-bounded-batch-queue-in-your-terminal-84.mdA Bounded Batch Queue in Your Terminal
Build a Python batch runner with bounded concurrency, backpressure, retries, and JSONL results through Fluxo.
-
rotate-fluxo-keys-during-a-kubernetes-rollout-83.mdRotate Fluxo Keys During a Kubernetes Rollout
Validate a replacement Fluxo API key, roll Kubernetes pods without dropping traffic, and revoke the old key only after verification.
-
a-terminal-git-helper-with-bounded-function-calls-82.mdA Terminal Git Helper With Bounded Function Calls
Build a small Git assistant with the OpenAI SDK, validated tool arguments, bounded execution, and a complete function-calling loop.
-
debugging-an-ai-api-401-from-your-terminal-81.mdDebugging an AI API 401 from Your Terminal
Trace an AI API 401 from shell variables to SDK configuration, with copyable commands that isolate credentials, URLs, and routing.
-
terminal-429s-exponential-backoff-with-curl-80.mdTerminal 429s: Exponential Backoff with curl
Debug AI API rate limits, honor Retry-After, and add bounded exponential retries with copyable curl commands.
-
stream-ai-tokens-to-your-terminal-with-nodejs-sse-79.mdStream AI Tokens to Your Terminal with Node.js SSE
Build a Node.js SSE endpoint, stream model output through Fluxo, and test incremental responses with curl using exact commands.
Sign up in a minute and get credit to try every model mentioned here. Sign up →