DEV Community

Dakota Huang profile picture

Dakota Huang

Building things with Python, Go, and JavaScript. Love automating everything.

Location Seattle, WA Joined Joined on 
A 429 Is a Signal, Not a Suggestion: Build a Retry Ladder for Free Endpoints

A 429 Is a Signal, Not a Suggestion: Build a Retry Ladder for Free Endpoints

Comments
6 min read
A/B Test Prompts Without a Feature Flag: A Free-Server Workflow

A/B Test Prompts Without a Feature Flag: A Free-Server Workflow

Comments
3 min read
Quota Forensics: A Token Ledger for Free Model Endpoints

Quota Forensics: A Token Ledger for Free Model Endpoints

Comments
4 min read
Real Traffic Is the Only Benchmark: Replay Requests Before You Swap Models

Real Traffic Is the Only Benchmark: Replay Requests Before You Swap Models

Comments
5 min read
A Quota Ledger: Track, Predict, and Cap Your Free Model Allowance

A Quota Ledger: Track, Predict, and Cap Your Free Model Allowance

Comments
6 min read
What Did the Model Decide? A Decision Log Tutorial for a Free Server

What Did the Model Decide? A Decision Log Tutorial for a Free Server

Comments
4 min read
Benchmarks Belong on a Server: A Deploy-and-Prove Tutorial

Benchmarks Belong on a Server: A Deploy-and-Prove Tutorial

Comments
5 min read
Five Gates to a Public Model Endpoint: A Deployment Tutorial

Five Gates to a Public Model Endpoint: A Deployment Tutorial

Comments
5 min read
Zero to Verified: A Free Model API on a Free Server, Stage by Stage

Zero to Verified: A Free Model API on a Free Server, Stage by Stage

Comments
6 min read
Free Servers Kill Processes. Make Your Model Jobs Survive.

Free Servers Kill Processes. Make Your Model Jobs Survive.

Comments
5 min read
Cache the Canonical Form: A Normalizing Cache for Metered Model Calls

Cache the Canonical Form: A Normalizing Cache for Metered Model Calls

Comments
5 min read
From Localhost to a Real Host: Deploy a Model Gateway in 6 Verified Steps

From Localhost to a Real Host: Deploy a Model Gateway in 6 Verified Steps

Comments
5 min read
The Cheapest Request Is the One You Never Send: A Semantic Cache for Rate-Limited APIs

The Cheapest Request Is the One You Never Send: A Semantic Cache for Rate-Limited APIs

Comments
6 min read
The Same Prompt Twice Is Double Quota: A Verifiable Cache for Free Model Endpoints

The Same Prompt Twice Is Double Quota: A Verifiable Cache for Free Model Endpoints

Comments
7 min read
From Prompt to Fixture: A Reliable Test Data Generator on a Free Model Endpoint

From Prompt to Fixture: A Reliable Test Data Generator on a Free Model Endpoint

Comments
3 min read
Don't Throttle, Schedule: A Priority Queue for Shared Model Quota

Don't Throttle, Schedule: A Priority Queue for Shared Model Quota

Comments
5 min read
Localhost Is Not a Server: A 3-Stage Probe for Free Model Endpoints

Localhost Is Not a Server: A 3-Stage Probe for Free Model Endpoints

Comments
4 min read
The 2 AM Incident: When a Free Model Endpoint Returned 200 and Lied

The 2 AM Incident: When a Free Model Endpoint Returned 200 and Lied

Comments 1
3 min read
Free Model Endpoints Rate-Limit You. Build a Retry Layer You Can Verify.

Free Model Endpoints Rate-Limit You. Build a Retry Layer You Can Verify.

Comments
4 min read
Free Model Endpoints Fail Quietly: A Self-Hosted Health Log in 40 Lines

Free Model Endpoints Fail Quietly: A Self-Hosted Health Log in 40 Lines

Comments
5 min read
Contract-Route Free Model Endpoints Before You Let Them Touch Real Requests

Contract-Route Free Model Endpoints Before You Let Them Touch Real Requests

Comments
5 min read
Build a Local Redacting Proxy for Free Model APIs in Under 60 Lines

Build a Local Redacting Proxy for Free Model APIs in Under 60 Lines

Comments
5 min read
Run Prompt Changes Like Code: A Tiny Regression Suite on Free Model Endpoints

Run Prompt Changes Like Code: A Tiny Regression Suite on Free Model Endpoints

Comments
4 min read
Shadow-Test Free Model Endpoint Changes Before You Rely on Them

Shadow-Test Free Model Endpoint Changes Before You Rely on Them

Comments
4 min read
Hash-Chain Your Free Model Responses Before You Trust Them

Hash-Chain Your Free Model Responses Before You Trust Them

Comments
5 min read
A Valid JSON Tool Call Can Still Delete the Wrong Row: Add an Argument Contract Before Dispatch

A Valid JSON Tool Call Can Still Delete the Wrong Row: Add an Argument Contract Before Dispatch

Comments
5 min read
Retry the Request, Not the Prompt: An Error Taxonomy for Free Coding Models

Retry the Request, Not the Prompt: An Error Taxonomy for Free Coding Models

Comments
3 min read
A Valid JSON Tool Call Can Still Delete the Wrong Row: Add an Argument Contract Before Dispatch

A Valid JSON Tool Call Can Still Delete the Wrong Row: Add an Argument Contract Before Dispatch

Comments
5 min read
200 OK Is Not a Model Answer: Triage Silent Empty Responses from Free Endpoints

200 OK Is Not a Model Answer: Triage Silent Empty Responses from Free Endpoints

Comments
3 min read
Free Models Love YAML: Add a Schema Diff Before You Merge Their Patches

Free Models Love YAML: Add a Schema Diff Before You Merge Their Patches

Comments
4 min read
Model-Generated GitHub Actions Need a Permission Gate

Model-Generated GitHub Actions Need a Permission Gate

Comments
3 min read
A Free Model Server Is Not a Sandbox: Budget Time, Bytes, and Redirects Before You Trust Output

A Free Model Server Is Not a Sandbox: Budget Time, Bytes, and Redirects Before You Trust Output

Comments
4 min read
Free Model Outputs Need a Schema Gate, Not Just a JSON Parse

Free Model Outputs Need a Schema Gate, Not Just a JSON Parse

Comments
3 min read
Before You Add Workers to a Free Model API, Sweep Its Effective Concurrency

Before You Add Workers to a Free Model API, Sweep Its Effective Concurrency

Comments
4 min read
Make Free Model Requests Idempotent Before You Add Retries

Make Free Model Requests Idempotent Before You Add Retries

Comments
4 min read
Model-Generated Setup Scripts: Lint the Egress Before You Run Them

Model-Generated Setup Scripts: Lint the Egress Before You Run Them

Comments
5 min read
Exit Code Zero Is Not a Clean Audit: Diff the Process Tree Before You Trust a Free Model's Shell

Exit Code Zero Is Not a Clean Audit: Diff the Process Tree Before You Trust a Free Model's Shell

Comments
3 min read
A Default-Deny Shell Is Not a Default-Deny Network: Probe What a Free Model Server Can Reach

A Default-Deny Shell Is Not a Default-Deny Network: Probe What a Free Model Server Can Reach

Comments
3 min read
A Receipt Checkpoint for Free Model Outputs Before They Become Shell Commands

A Receipt Checkpoint for Free Model Outputs Before They Become Shell Commands

Comments
4 min read
Exit Code Zero Hides New Processes: Diff the Process Table After a Model-Generated Command

Exit Code Zero Hides New Processes: Diff the Process Table After a Model-Generated Command

Comments
4 min read
Before You Automate a Free Model Endpoint, Measure Rate, Shape, Drift, and Budget

Before You Automate a Free Model Endpoint, Measure Rate, Shape, Drift, and Budget

Comments
4 min read
Free Server, Untrusted Model: Diff the Filesystem Before You Trust the Exit Code

Free Server, Untrusted Model: Diff the Filesystem Before You Trust the Exit Code

Comments
3 min read
Free Servers Should Never See Your Real Credentials

Free Servers Should Never See Your Real Credentials

Comments
4 min read
A Permission Matrix Before Your Free Model Touches the Shell

A Permission Matrix Before Your Free Model Touches the Shell

Comments
4 min read
A Contract Probe for Free Model Endpoints That Drift

A Contract Probe for Free Model Endpoints That Drift

Comments
4 min read
A Streaming Response Is Not a Payload: Buffer JSON Before You Validate

A Streaming Response Is Not a Payload: Buffer JSON Before You Validate

Comments 1
4 min read
A Read-Only Gate for Model-Generated SQL on Free Compute

A Read-Only Gate for Model-Generated SQL on Free Compute

Comments
4 min read
Draft PR Descriptions From Diffs Without Giving the Model Write Access

Draft PR Descriptions From Diffs Without Giving the Model Write Access

Comments
3 min read
Split Low-Stakes Prompts From Code Changes Before You Call a Free Model

Split Low-Stakes Prompts From Code Changes Before You Call a Free Model

Comments
3 min read
Pin Free Inference Before You Trust It: A Golden Hash Gate

Pin Free Inference Before You Trust It: A Golden Hash Gate

Comments
4 min read
A Free Model Endpoint Needs a Cold-Start Budget

A Free Model Endpoint Needs a Cold-Start Budget

Comments
2 min read
Gate the Patch, Not the Model: A Three-Check Loop for Untrusted Code Outputs

Gate the Patch, Not the Model: A Three-Check Loop for Untrusted Code Outputs

Comments 1
5 min read
A Noise Budget for Free Model Evals: Live vs. Replay

A Noise Budget for Free Model Evals: Live vs. Replay

Comments
5 min read
A Deploy Smoke Test for Model-Generated Code

A Deploy Smoke Test for Model-Generated Code

Comments
4 min read
Don't Merge AI Patches Blind: A Git Worktree Replay Gate

Don't Merge AI Patches Blind: A Git Worktree Replay Gate

Comments
5 min read
Same Model, Different Server: Fingerprint Your Eval Environment Before You Trust the Result

Same Model, Different Server: Fingerprint Your Eval Environment Before You Trust the Result

Comments
4 min read
Probe the Endpoint Before You Probe the Model

Probe the Endpoint Before You Probe the Model

Comments
5 min read
A Failure Triage Script for Free Coding Models: Prompt, Model, or Wrapper?

A Failure Triage Script for Free Coding Models: Prompt, Model, or Wrapper?

Comments
4 min read
From Six Questions to a Script: My 30-Minute Eval Harness for Every New Model Release

From Six Questions to a Script: My 30-Minute Eval Harness for Every New Model Release

Comments
6 min read
I Stopped Trusting Demo Prompts: A Repeatable Smoke Test for Free Coding Models

I Stopped Trusting Demo Prompts: A Repeatable Smoke Test for Free Coding Models

Comments
5 min read
loading...