Tag feed

#llm

6,692 posts317 followers

Explore Hashnode

Alternatives

Trending tags this week

Ccode&cosmoscodeandcosmos.hashnode.dev3h ago · 13 min read

The Developer's Guide to AI Tokenomics: Why Your Coding Agent Can Burn Through Credits in Minutes

You never came close to your context limit, yet your team's AI credit balance dropped sharply. Here's the real math behind agent loops, hidden context, reasoning tokens and prompt caching. A develope

0

SMShubham Moretheshubhammore.hashnode.dev9h ago · 12 min read

Building with LLMs? Understand Tokens, API Costs, and How to Avoid Unexpected Bills

Introduction: The Part of LLM Development We Often Overlook When we start building applications with Large Language Models (LLMs), we're usually excited about what we can create. We experiment with di

0

江江流石jliushi.hashnode.dev9h ago · 5 min read

Routing Claude Code (and Codex) Through an Anthropic-Compatible Relay: Env Vars, settings.json and Status-Code Debugging

TL;DR export ANTHROPIC_BASE_URL="https://<relay-base-url>" # no trailing /v1 export ANTHROPIC_AUTH_TOKEN="<relay-key>" # sent as Bearer export ANTHROPIC_MODEL="claude-opus-5"

0

TTech-Aarvamtechaarvam.hashnode.dev5h ago · 3 min read

Our in-house agent router: Jev vs Kev-4B vs bge-small

In our (currently) in-house business automation agent, we route prompts to the right agent. This is done using an agent router. Take-away: before a prompt reaches the active agent, two small local mod

0

UKUsman Khanctousman.hashnode.dev5h ago · 5 min read

Server-Sent Events (SSE) Backpressure Breaks: Memory Leaks and Socket Starvation in High-Throughput LLM Pipelines

Streaming LLM completions over Server-Sent Events (SSE) looks trivial in sandbox demos, but unhandled backpressure under real-world network conditions can destroy backend application servers. When slo

0

SRSai Ram Kothasairamkotha.hashnode.dev14h ago · 7 min read

Your MCP Servers Bill You on Every Request: The Tool-Definition Tax

You connect a few MCP servers and your agent gains dozens of abilities. What the connection also adds is a block of text that rides along in the prompt of every single request, whether the agent uses

0

NSNeeraj Sharmaaxionry.hashnode.dev7h ago · 9 min read

DeepSeek V4.1 Flash agent loops: your cache-hit rate is the bill

The post behind this week's Hacker News thread, "Why isn't the industry freaking out about DeepSeek 4.1 Flash?", credits DeepSeek's smaller KV cache for keeping all-day sessions under a dollar. I want

0

HKHarshavardhan Katkamharshakatkam.dev1d ago · 9 min read

Day 01: Why your first agent should be 40 lines, not a framework

The first agent I ever got working, I did not understand. I had installed a framework, pasted eleven lines out of its quickstart, and watched a correct answer appear. Then I changed the prompt slightl

1

A

DDVARAdvarahq.hashnode.dev18h ago · 4 min read

A Policy Rule Describes the Violation, Not the Goal: LLM Policy Enforcement in Practice

A YAML file full of model rules is documentation until something enforces it on every request. LLM policy as code is useful when three things travel together: a reviewable rule, one shared enforcement

0

AKAyesha Khanayeshakoder.hashnode.dev1d ago · 7 min read

Believable Is Not Correct

Our owner wants an AI agent in the app. It should understand the regional language, take voice notes, cope with different dialects, and do real work inside the product. It's an exciting request, and I

0

#llm

Search Hashnode

#llm

Explore Hashnode

Trending tags this week

The Developer's Guide to AI Tokenomics: Why Your Coding Agent Can Burn Through Credits in Minutes

Building with LLMs? Understand Tokens, API Costs, and How to Avoid Unexpected Bills

Routing Claude Code (and Codex) Through an Anthropic-Compatible Relay: Env Vars, settings.json and Status-Code Debugging

Our in-house agent router: Jev vs Kev-4B vs bge-small

Server-Sent Events (SSE) Backpressure Breaks: Memory Leaks and Socket Starvation in High-Throughput LLM Pipelines

Your MCP Servers Bill You on Every Request: The Tool-Definition Tax

DeepSeek V4.1 Flash agent loops: your cache-hit rate is the bill

Day 01: Why your first agent should be 40 lines, not a framework

A Policy Rule Describes the Violation, Not the Goal: LLM Policy Enforcement in Practice

Believable Is Not Correct