Getting started¶

Installation¶

Requires Python 3.12+.

Install from source with uv:

git clone https://github.com/aaronwolen/transcript-indexer.git
cd transcript-indexer
uv sync

Basic usage¶

Point txi at a directory of transcripts (defaults to ~/transcripts), sync the index, then search.

# One-off parse of a single file as a sanity check
txi parse path/to/transcript.txt

# Walk the corpus and bring the index into sync with the filesystem
# (this also computes embeddings for new/changed conversations).
txi sync

# Full-text, semantic, or hybrid search
txi search fts "release planning"
txi search semantic "how should we sequence the launch?"
txi search hybrid "release planning"

Use --help on any command or subcommand to see available options:

txi --help
txi sync --help
txi search --help
txi people --help

Command overview¶

Command	Description
`txi parse`	Parse a single transcript file and print it as JSON
`txi sync`	Walk the corpus and bring the index into sync with the filesystem
`txi embed`	Run the embedding pipeline against indexed conversations
`txi serve`	Launch the local web dashboard
`txi check`	Validate transcript files against a specific reader format
`txi stats`	Print corpus statistics
`txi search fts`	Full-text search across transcript turns
`txi search semantic`	Vector (semantic) search across chunk embeddings
`txi search hybrid`	Hybrid FTS + semantic search merged via RRF
`txi conversations`	List conversations matching metadata filters
`txi conversation`	Show a single conversation: header and turns
`txi people list`	List every canonical person and their aliases
`txi people add`	Add a new canonical person with optional aliases
`txi people alias`	Add or remove an alias for an existing person
`txi people merge`	Merge two people, reassigning all foreign keys
`txi people unknown`	List raw values not yet resolved to a person
`txi people resolve`	Re-run alias resolution without re-reading sources
`txi config path`	Print the path of the active config file
`txi config view`	Print the contents of the active config file
`txi config validate`	Validate the config file's TOML syntax and field types
`txi db path`	Print the path of the SQLite index database
`txi db reset`	Delete the SQLite index database so the next sync starts fresh