NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #1939 most downloaded on PyPI
Bindings for the Low-level Guidance (llguidance) Rust library for use within Guidance
Last release 5 days ago
30 Sep 2026
Release timing varies
gaps range from 8 days to 3 months
Most releases are documented
notes for 52 of the last 60 stable releases
Nothing withdrawn
no release was ever pulled
2 years old
92 releases · first in 2024
Expose cooperative cancellation on TokenParser `#387`
#387Add opt-in cancellation for abandoned matcher computations `#382`; fixes `#261`
#382; fixes #261
#376
#378; fixes #226
#372; fixes #206
#379One column per month.
Allow general Unicode escapes in JSON strings and object keys `#374`
#374\xHH escapes in Lark string literals #354
#370#368
#369
#352 #353max_tokens=0 as an empty string #356
multipleOf #357
Fix validate_tokens for multiple EOS tokens `#347`
Refactor decoder detection to use enum matching with recursive traversal `#331`
#331#344
Fix decoder type detection for Sequence containing ByteLevel `#325`
#335
#325
Nothing published for this version
Add support for multiple EOS tokens `#305`
#305#308get_capture and get_captures API to Python #289
Nothing published for this version
harden maturin release; add skip-existing so we can upload failures
harden maturin release; add skip-existing so we can upload failures
#288
#285; fixes #256
#286
#287
#283#282
Mask caching to improve performance, especially for lazy lexemes `#279`
[FIX] Coerce Metaspace prepend_scheme to never `#272`
Metaspace prepend_scheme to never #272#264
#273#271
#270#269
#260
Update pyo3 dependency to latest version to support python 3.14t `#255`
#255
#249
#25230d1515
change semantics of JSON separators (now regex patterns instead of text literals) `#246`
support multithreaded compute bitmask for speculative decoding `#225`
add flag in ParserLimits to disable verbose errors `#227`
#227prevent error state when calling try_consume_tokens after parser is stopped `#213`; fixes `#211`
disable hf tokenizer truncation and padding `#205`; fixes `#1322`
fix: tokenizers normalizers sequence api changed `#195`
This is identical to 0.7.30, but indicates intended stability and from now on we'll try to follow semver.
This is identical to 0.7.30, but indicates intended stability and from now on we'll try to follow semver.
allow for tokens up to ~2k bytes; fixes #188 `#188`
- cargo fmt
add toktrie_tiktoken and llguidance.tiktoken.lltokenizer_from_encoding `#154`
add support for & and ~ in lark regexes `96fcee3`
add parse_special=False to tokenize_str/bytes() in python `#181`
#181add the sentinel token hack, fixes #180 `#180`
#180native llama.cpp tokenizer support `#179`
include parser state in errors `82e34da`
use fancy-regex instead of onig as tokenizers regex library `#172`
#172
c89e386, thank you @nteodosio for initial PR!fix a numeric token bug `1f59edf`
1f59edfapply x-guidance also in %json{} `2627891`
support for min/maxProperties in JSON Schema `#168`
fix special token tokenization `ae7870f`
ae7870fsupport for patternProperties in JSON schema `#167`
support %lark { ... } syntax for nested grammars `#157`
#157#158929b13fbb228cb78fb32fca73c2aexpose LLParserLimits in Python API `598dc8f`
598dc8fuse factory in C FFI (otherwise slicer was not used)
#113add StructTag python API; fixes `#146`
add llg_matcher_*() functions to C interface `#145`
improve Python LLMatcher.validate_grammar() `6b5f5ed`
LLMatcher.validate_grammar() 6b5f5edNothing published for this version
Nothing published for this version
Stabilize JSON schema property order `#134`
#134add toktrie_hf_downloader crate `11bea00`
- fix gbnf parsing `e5828b8`
e5828b8add LLMatcher.validate_grammar(); make it never raise `0f8ec60`
don't go into error state on final EOS `1f0f21d`
1f0f21dadd LLMatcher interface in python
remove JSON-based grammar serialization `#133`
#133fix https://github.com/guidance-ai/guidance/issues/1131 - backtracking+prompt healing `#1131`
Nothing published for this version
[JSON] "x-guidance" JsonCompileOptions `#130`
#130allow lexer to produce alternative lexemes `#124`
fix #122: captures with nullable symbols `#122`
#122Extend Token::Number to match floats in scientific notation `#121`
update referencing to 0.29.0 `#118`
Nothing published for this version
Nothing published for this version
Nothing published for this version
Your coding agent can read these notes before it upgrades. Set up the MCP server →