Skip to content
Veer
Search
Ctrl
K
Cancel
GitHub
Select theme
Dark
Light
Auto
Menu
Download
Playground
Start here
Installation
Quickstart
Why Veer?
Learn the background
Overview
1. How LLMs generate text
2. The KV cache
3. Serving systems
4. Agents and context editing
5. Number formats and quantization
6. Error analysis
7. Types, linearity, and static analysis
Glossary
Tutorial
1. Contexts
2. Forking and sharing
3. Structured generation
4. Editing contexts with splice
5. Speculation
6. Loops, functions, and budgets
7. Precision types
Reference
Syntax
Types
Linearity rules
Context operations
The planner
Precision layer
Builtins
CLI
Engines
Limitations and roadmap
Cookbook
Code review with sharing, editing, and speculation
Agent loop with KV editing
Tree-of-thought search
Many questions over one document
Precision types: choosing a 4-bit format
Hello, Veer
Research and motivation
Error index
Error index
V0001: invalid token
V0002: syntax error
V0100: unknown name
V0101: use of moved context
V0102: context is never consumed
V0103: branches disagree about a context
V0104: loop changes context state
V0105: context cannot be duplicated
V0106: model mismatch
V0110: type mismatch
V0111: wrong number of arguments
V0112: string bound exceeded
V0113: unbounded loop
V0114: recursion is not allowed
V0115: unknown field
V0116: structured output does not fit
V0117: unbounded string appended to a context
V0118: invalid declaration
V0119: duplicate definition
V0120: unknown span
V0121: duplicate span label
V0130: speculation must end in commit or rollback
V0131: outer context used inside speculation
V0132: statement not allowed here
V0200: KV memory budget exceeded
V0201: prefill budget exceeded
V0202: context length exceeded
V0300: tensor shape mismatch
V0301: accuracy bound violated
V0302: possible overflow in number format
V0303: unknown number format
W0210: splice forces re-prefill
GitHub
Select theme
Dark
Light
Auto
V0202: context length exceeded
Some context could grow longer than the model’s
max_ctx
.