Issues

• OpenAI provider sends max_tokens to gpt-5 and o-series models, which reject it

  google/langextract#598 · Filter by author pcbeingused333 opened on Oct 29, 2026

• Bug: max_char_buffer=0 or a negative value calls the model once per token (and crashes with debug=True)

  google/langextract#589 · Filter by author tonmoy-y opened on Oct 26, 2026

• Bug: extraction_passes=0 or a negative value silently returns no extractions

  google/langextract#587 · Filter by author tonmoy-y opened on Oct 26, 2026

• Bug: read_prompt_template_structured_from_file decodes UTF-8 prompt files with the locale encoding

  google/langextract#585 · Filter by author tonmoy-y opened on Oct 26, 2026

• Bug: fenced JSON parsing rejects triple backticks inside extraction strings

  google/langextract#581 · Filter by author BichengWang opened on Oct 23, 2026

• Bug: changing Document.text after extraction reuses stale tokens and produces incorrect exact matches

  google/langextract#580 · Filter by author BichengWang opened on Oct 21, 2026

• Bug: multiple visualize() outputs on one page overwrite each other's controls and attributes

  google/langextract#578 · Filter by author BichengWang opened on Oct 21, 2026
  1

• Bug: dict_to_annotated_document mutates its input and fails when reused

  google/langextract#576 · Filter by author BichengWang opened on Oct 19, 2026

• Add DemonRoute as a documented OpenAI-compatible endpoint

  google/langextract#575 · Filter by author haterfactory opened on Oct 19, 2026

• Bug: documents saved with no extractions cannot be loaded back (TypeError)

  google/langextract#574 · Filter by author LemonCondor opened on Oct 15, 2026
  2

• Request: Optional web source discovery and extraction example with Parallel Search MCP

  google/langextract#568 · Filter by author georgeatparallel opened on Oct 11, 2026

• Request: choose which document visualize() renders from a multi-document JSONL

  google/langextract#563 · Filter by author BichengWang opened on Oct 8, 2026
  1

• Bug: downloaded text ignores charsets and corrupts Unicode BOM encodings

  google/langextract#557 · Filter by author cangen185-sudo opened on Oct 5, 2026

• Request: default_headers for the OpenAI provider

  google/langextract#549 · Filter by author domondi1 opened on Oct 3, 2026

• Request: allow callers to filter or score chunks before inference

  google/langextract#545 · Filter by author reichaves opened on Sep 23, 2026
  1

• OrcaRouter provider plugin (langextract-provider-orcarouter)

  google/langextract#524 · Filter by author JinhaoSong322 opened on Aug 20, 2026

• Request: Built-in Anthropic (Claude) and xAI (Grok) providers

  google/langextract#522 · Filter by author jaideepkumar69-tech opened on Aug 20, 2026

• Bug: UnicodeEncodeError in create_provider_plugin.py on Windows (cp1252 console)

  google/langextract#518 · Filter by author jaideepkumar69-tech opened on Aug 18, 2026

• Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

  google/langextract#487 · Filter by author granjacours697-netizen opened on Jul 7, 2026

• Feature Request: Token Usage Tracking (or model metadata?) Support

  google/langextract#475 · Filter by author g-jensen opened on May 29, 2026
  1

• Performance: prompt building duplicates full few-shot preamble per prompt, causing O(batch_length × preamble_size) peak memory

  google/langextract#446 · Filter by author dan504512 opened on Apr 17, 2026
  2

• test: Add missing test coverage for chunking module edge cases

  google/langextract#430 · Filter by author vedssharma opened on Apr 1, 2026
  1

• Inference cost

  google/langextract#425 · Filter by author TheMrguiller opened on Mar 26, 2026
  1

• ResolverParsingError when using vLLM/lmdeploy as backend (Qwen3-32B-GPTQ)

  google/langextract#414 · Filter by author wangjiawen2013 opened on Mar 13, 2026
  2

• Build ML Model to Predict Public Theater Sessions

  google/langextract#376 · Filter by author BrDataRep opened on Feb 16, 2026