0.4.25 (04 April 2026)

  • Scanners: Distinguish thinking, thinking summary, and redacted thinking when rendering messages for scanners.
  • Scanners: Surround assistant messages with prefill=True in metadata with <prefill> tag when rendering messages for scanners.
  • Scanners: Match references across all fields in scan result.
  • Metrics: Filter out NaN metrics (occurs when all scan results are errors).
  • Metrics: Filter out non-numeric keys from scanner values when computing scanner metrics.
  • Timelines: Ensure that all events are loaded from reading timelines.
  • Timelines: Update to use new timeline branch scheme from inspect-ai.
  • Timelines: Add label_for_id option to message_numbering() function.
  • Modules: Rename async module to aio.
  • VS Code Integration: Fix regression displaying Scout View in VS Code.
  • Scout View: Show tooltip for array or object values which have been truncated.
  • Scout View: Improve citation matching in results view.

0.4.23 (25 March 2026)

  • Add timeline types to @scanner decorator overloads.
  • Scout View: Fix white circle over compaction markers.
  • Bugfix: Handle excluded input column in event expansion.
  • Bugfix: Always truncate error file on resume to prevent stale errors.

0.4.22 (20 March 2026)

  • Transcripts: Use events_data to reduce memory and storage requirements of events.
  • Summary: Count only positive values in resultset aggregation.
  • Summary: Reset summary state on new scan init to prevent accumulation across scans.
  • Recorder: Preserve error file on scan resume instead of truncating.

0.4.21 (16 March 2026)

  • Scanner as scorer: Forward timelines from eval sample to scanner if requested.
  • Scout View: Correctly display timelines in results view.
  • Scout View: Show dictionary-based metriecs in scanner sidebar.
  • Scout View: Bundle dist/ assets during packaging.

0.4.20 (16 March 2026)

  • LLM Scanner: Automatic transcript segmentation by context window with configurable compaction handling.
  • LLM Scanner: AnswerMultiLabel(allow_none=True) lets the model respond with ANSWER: NONE when no labels apply.
  • Scanner Tools: generate_answer() automatically retries with format feedback when the model’s response can’t be parsed.
  • Scanner Tools: message_numbering(), scanner_prompt(), generate_answer(), and parse_answer() functions for building custom scanners with fine-grained control.
  • Scanner Tools: ResultReducer for reducing results from multiple transcript segments into single results, with built-in majority and LLM-based reducers.
  • Scanner Tools: transcript_messages(), segment_messages(), and span_messages() functions for extracting and segmenting transcript messages.
  • Transcript DB: claude_code() source for importing transcripts from Claude Code session logs. Supports filtering by project, session, and time range, session merging, and image extraction.
  • CLI: scout import command for importing transcripts from registered sources into Scout projects.
  • Serialization: Use pa.large_string for string types to support larger column/file sizes.
  • Multiprocessing: Improve handling of model instances with multiprocessing serialization.
  • Transcripts: unthin target and and add scores from sample JSON.
  • Transcripts: Set row group size to 25 (specify as rows not bytes).
  • Transcripts: Address DuckDB 1.5 compatibility issue w/ mixed type CASE expressions.
  • Transcripts: Switch over to async ZIP modules (async_zip, zip_common, compression, compression_transcoding, async_bytes_reader) that have migrated to inspect_ai.
  • Transcripts: Remove parquet encryption (not used + issues w/ DuckDB 1.5).
  • Transcripts: Ensure that all documented schema columns exist when running transcript queries.
  • Observe: Prevent transcript index staleness/warning from occurring when running parallel observe contexts.
  • Scout View: Properly sort scanner results using the value type.
  • Scout View: Enable minification and caching of view static assets.
  • Scout View: Fix issue rendering transcript events when showing the validation panel.
  • Scout View: Add ‘None’ option to column chooser to unselect all columns.
  • Bugfix: Fix early-exit bug the failed to unthin sample_metadata

0.4.19 (17 February 2026)

  • LLM Scanner: Store model stop_reason result metadata.
  • Scout View: Fix incorrect behavior when attempting to view scanner results with result sets.

0.4.18 (14 February 2026)

  • Bugfix: Fix token counting when a single worker task processes multiple scans sequentially.

0.4.17 (13 February 2026)

  • Transcript DB: Improve Anthropic and Google tool call capture for pheonix() transcript source.

0.4.16 (13 February 2026)

  • Transcript DB: phoenix() transcript source for importing transcripts from Arize Phoenix.

0.4.15 (12 February 2026)

  • Scout View: Support for displaying cost limits.

0.4.14 (12 February 2026)

  • Scout View: Improve display of large dictionary scan values.
  • Enable customization of the scan buffer directory via SCOUT_SCANBUFFER_DIR environment variable.
  • Bugfix: Fix answer parsing when LLM echoes “Answer:” in reasoning before the actual answer marker.

0.4.13 (11 February 2026)

  • Validation: Support for applying validation labels in Scout View.
  • Compatibility with latest release of the Inspect VS Code Extension.

0.4.12 (09 February 2026)

  • Scoring: Apply content filter for scanners when using them as Inspect scorers.
  • Add SampleMetadata class for typed access to Inspect eval log metadata fields.
  • Add support for Inspect CompactionEvent and display of native compaction data from OpenAI and Anthropic.
  • Add “store” event to EventType enumeration.
  • View Server: Add optimized /transcripts/{dir}/{id}/info and /transcripts/{dir}/{id}/messages-events endpoints for fetching transcript data. The messages-events endpoint streams raw (potentially compressed) JSON for improved performance.
  • Bugfix: Eliminate problem with stale transcript status when deleting validation cases.

0.4.11 (29 January 2026)

  • Projects: Always read scout.local.yaml even if there is no scout.yaml file.
  • Projects: Always apply project level filter to scans (AND combine with scan filters).
  • Validation: Label validation is now binary: validate true if the label is present with a truthy value; validate false if the label is not present or has only falsey values.
  • Scan Results: Add exclude_columns parameter for reading parquet reuslts to optionally reduce memory usage.
  • Scan Results: Pre-fetch optimization for S3/remote parquet files.
  • Transcript DB: observe() decorator/context manager for writing transcripts based on observed LLM generations.
  • Transcript DB: langsmith() and logfire() transcript sources for importing transcripts from LLM observability systems.

0.4.10 (21 January 2026)

  • Scanning: Implement significant scanning performance improvement when scanning eval logs and events are unneeded.
  • Scan config: Set ‘model’ to None if no model is specified.
  • Scan config: Deprecate use of environment variables for config (in favor of project config).
  • Scout View: Move ‘Project’ UI button to main activity bar.

0.4.9 (20 January 2026)

  • Bugfix: Don’t check index coverage when running with an active limit or other query filter.
  • Bugfix: Correctly normalize relative database file paths.

0.4.8 (18 January 2026)

  • Validation: Add support for defining and using named splits (e.g. ‘dev’, ‘test’) for validation data.
  • Validation: Add support for specifying per-case predicates within validation data.
  • Add message_count as standard transcript metadata field.
  • Bugfix: Correct async generator cleanup in AsyncBytesReader.

0.4.7 (16 January 2026)

  • Scout View: Editing UI for project settings.
  • Scanning: Apply model config when resuming scans.
  • Transcript DB: Warn when there is no index or the index is out of date.
  • Validation: Improve error messages and documentation; deprecate (with warning) headerless CSVs.
  • CLI: Add scout --version to print current scout version.

0.4.6 (08 January 2026)

  • LLM Scanner: Add value_to_float option to for converting model reported values to numeric.
  • Projects: Support local project config in scout.local.yaml.
  • Transcripts: Enable use of SQL for specifying filters.
  • Transcripts: Add filter field to scan job and project config.
  • Scanning: Add tool_callers() helper function for mapping tool_call_id to assistant message.
  • Rename --results option to --scans.
  • Bugfix: Avoid UnboundLocalError by importing Inspect AI batch reporting functions directly.
  • Bugfix: Fix issue with reading large numbers of rows from transcript database.

0.4.5 (03 January 2026)

  • Projects for centrally managing scanning configuration.
  • Grep Scanner for pattern-based scanning of transcripts.
  • Scanning: Add --dry-run option for previewing scanner counts.
  • Scanners: Fixup type annotations in @scanner decorator.
  • Display: Add text progress and metrics support to display="log" mode.
  • Transcript DB: Don’t remove orphaned data files.
  • Transcript DB: Handle duplicate transcript ids while indexing.
  • Transcript DB: Remove redundant TranscriptSource class (covered by AsyncIterable already).
  • Bugfix: Store relative paths in transcript database index.

0.4.4 (24 December 2025)

  • Bugfix: Restore compatibility with v1 view server API.

0.4.3 (22 December 2025)

  • Add more standard fields to the transcript database schema.
  • Add an optional index to transcript database for higher performance queries on very large databases.
  • Add as_json option to messages_as_str for JSON output format.
  • Persist transcripts.where() clauses as part of scan specification.
  • Bugfix: Fix LLM answer parsing for decimals, negatives, and markdown.

0.4.2 (14 December 2025)

  • Scanning: Switch from dill to cloudpickle and restore default max_processes to 4 after resolving multiprocessing serialization issues.
  • Track scan job completion status explicitly in the filesystem (vs. merely looking at whether buffer dir exists).
  • Transcript databases: Include all .parquet files in directory (don’t require transcripts_ prefix).

0.4.1 (12 December 2025)

  • Scan jobs: Correct resolution order for options (CLI, then scanjob config, then environment variables).
  • Scanning: Restore default max_processes to 1 while we resolve some multiprocessing serialization issues.

0.4.0 (11 December 2025)

  • Initial release.