DEV Community

Quinn Li profile picture

Quinn Li

Building and writing about AI-powered developer tools. Obsessed with clean code and useful automation.

Location Singapore Joined Joined on 
Count Minutes and Tokens in the Same Ledger

Count Minutes and Tokens in the Same Ledger

Comments
4 min read
The Free Tier Is an Instrument, Not a Budget

The Free Tier Is an Instrument, Not a Budget

Comments
4 min read
Do the Deadline Math Before You Enqueue

Do the Deadline Math Before You Enqueue

Comments
5 min read
Queue Time Is a Line Item

Queue Time Is a Line Item

Comments
5 min read
Retries Are Rent

Retries Are Rent

Comments
4 min read
Price the Run, Not the Token

Price the Run, Not the Token

Comments
5 min read
Book the Partial Before You Retry

Book the Partial Before You Retry

Comments
7 min read
Write the Prefix Down Before the Slot Dies

Write the Prefix Down Before the Slot Dies

Comments
6 min read
Free Tokens Do Not Pay the Clock

Free Tokens Do Not Pay the Clock

Comments
7 min read
Stamp a Deadline Before You Pick a Lane

Stamp a Deadline Before You Pick a Lane

Comments
7 min read
Refuse the Round You Have Not Sent

Refuse the Round You Have Not Sent

Comments
7 min read
No Socket Without a Fresh Envelope

No Socket Without a Fresh Envelope

Comments
7 min read
Don't Enqueue What You Cannot Reserve

Don't Enqueue What You Cannot Reserve

Comments
6 min read
Don't Put a Retry Loop on Free Capacity

Don't Put a Retry Loop on Free Capacity

Comments
7 min read
Don't Rehearse a Bill on Gift Capacity

Don't Rehearse a Bill on Gift Capacity

Comments
6 min read
Killed Work Still Hits the Meter

Killed Work Still Hits the Meter

Comments
7 min read
Free Capacity Is Spot. Committed Work Is Not.

Free Capacity Is Spot. Committed Work Is Not.

1
Comments 1
5 min read
Time in Queue Is Not Free

Time in Queue Is Not Free

Comments
7 min read
Give Free Capacity a Deadline

Give Free Capacity a Deadline

Comments
6 min read
The Call Graph Is What You Owe

The Call Graph Is What You Owe

1
Comments
6 min read
Retries Are How Cheap Jobs Get Expensive

Retries Are How Cheap Jobs Get Expensive

Comments
6 min read
Count Tokens per Success, Not per Request

Count Tokens per Success, Not per Request

Comments
5 min read
The Transcript Is the Bill

The Transcript Is the Bill

Comments 1
6 min read
You Keep Paying for the Same Prompt

You Keep Paying for the Same Prompt

Comments 2
7 min read
Blocked Time Makes Free Capacity the Wrong Bet

Blocked Time Makes Free Capacity the Wrong Bet

Comments
7 min read
Deadline Work Does Not Belong on a Free Lane

Deadline Work Does Not Belong on a Free Lane

Comments
7 min read
Charge Wait Time to the Same Job

Charge Wait Time to the Same Job

Comments
7 min read
Every Extra Hop Buys Another Queue Ticket

Every Extra Hop Buys Another Queue Ticket

Comments
7 min read
Budget the Retry Path, Not the Happy Path

Budget the Retry Path, Not the Happy Path

Comments
6 min read
The Second Call Is a Different Job

The Second Call Is a Different Job

Comments
8 min read
Write the Kill Envelope Before the First Prompt

Write the Kill Envelope Before the First Prompt

Comments
7 min read
3 A.M. Token Leak: An Autopsy of an Agent Loop

3 A.M. Token Leak: An Autopsy of an Agent Loop

Comments
5 min read
Score Your Model Access Decision Before You Argue

Score Your Model Access Decision Before You Argue

Comments
4 min read
If the Job Has a Deadline, Spare Capacity Is the Wrong Bet

If the Job Has a Deadline, Spare Capacity Is the Wrong Bet

Comments
5 min read
Your Hourly Rate Makes Free Tokens Expensive

Your Hourly Rate Makes Free Tokens Expensive

Comments
7 min read
Free Model Access Needs an Exit Plan: How to Test LLM Endpoints Without Betting Your Pipeline

Free Model Access Needs an Exit Plan: How to Test LLM Endpoints Without Betting Your Pipeline

Comments
4 min read
The Retry Tax: When Free AI Capacity Stops Being Free

The Retry Tax: When Free AI Capacity Stops Being Free

Comments
3 min read
Free Tokens Are a Tool, Not a Promise: Measure Before You Build

Free Tokens Are a Tool, Not a Promise: Measure Before You Build

1
Comments
4 min read
Cheap Tokens, Expensive Waiting Rooms

Cheap Tokens, Expensive Waiting Rooms

Comments
4 min read
Free Tokens Are a Queue, Not a Wallet

Free Tokens Are a Queue, Not a Wallet

1
Comments
5 min read
Free Tokens, Real Queues: Measure What Your LLM Calls Actually Cost

Free Tokens, Real Queues: Measure What Your LLM Calls Actually Cost

Comments
4 min read
Chaos for the Cost-Conscious: Fault-Injecting Your LLM Pipeline on Free Hardware

Chaos for the Cost-Conscious: Fault-Injecting Your LLM Pipeline on Free Hardware

Comments 1
4 min read
Free, Paid, or Self-Hosted: Run the Decision Script

Free, Paid, or Self-Hosted: Run the Decision Script

Comments
4 min read
Routing Around Rate Limits: A Free-Tier Model Proxy in 100 Lines

Routing Around Rate Limits: A Free-Tier Model Proxy in 100 Lines

Comments
4 min read
The Free-Tier Trap: A Framework for Choosing Model Access

The Free-Tier Trap: A Framework for Choosing Model Access

Comments
5 min read
A Free Server Is Enough to Test a New Model Before You Trust It

A Free Server Is Enough to Test a New Model Before You Trust It

Comments
3 min read
Reproducible Evaluation Harness for New Model Releases: MiniMax H3

Reproducible Evaluation Harness for New Model Releases: MiniMax H3

Comments
5 min read
A Reproducible Tool-Call Gatekeeper for AI Agents

A Reproducible Tool-Call Gatekeeper for AI Agents

Comments
4 min read
loading...