Configuration

All settings can be persisted via imprint config instead of setting environment variables. Settings are stored in data/config.json (gitignored).

Precedence: env var > config.json > hardcoded default. Environment variables always win, so you can override config.json for one-off runs.

# Switch to a different embedding model
imprint config set model.name nomic-ai/nomic-embed-text-v2-moe
imprint config set model.dim 768
imprint config set model.seq_length 512

# Use local Ollama for LLM tagging
imprint config set tagger.llm true
imprint config set tagger.llm_provider ollama
imprint config set tagger.llm_model llama3.2

# Or fully in-process via llama-cpp (no server to run)
imprint config set tagger.llm true
imprint config set tagger.llm_provider local

# Custom Qdrant server
imprint config set qdrant.host 192.168.1.50
imprint config set qdrant.no_spawn true

# See what's changed
imprint config

# Show one setting + where the current value came from
imprint config get model.device

# One-off env override (doesn't persist)
IMPRINT_DEVICE=gpu imprint ingest ~/code

# Remove one override
imprint config reset model.device

# Wipe all overrides
imprint config reset --all

imprint config groups settings visually by prefix (model.*, qdrant.*, chunker.*, tagger.*, tagger.local.*, ingest.*, chat.*, summarizer.*). Non-default values are highlighted by source (config.json cyan, env yellow).

All Settings

Key	Default	Description
Embedding model
model.name	onnx-community/embeddinggemma-300m-ONNX	HuggingFace embedding model repo
model.file	auto	ONNX model file (auto = pick by device)
model.device	auto	Compute device: auto / cpu / gpu
model.dim	768	Embedding vector dimension
model.seq_length	2048	Token cap per embed call
model.threads	4	CPU intra-op threads for ONNX
model.gpu_mem_mb	2048	VRAM cap for ORT CUDA arena (MB)
model.gpu_device	0	CUDA device ID
model.batch_size	0	Embedding batch size (0 = auto: 32 GPU, 16 CPU)
model.pooling	auto	Pooling strategy: auto / cls / mean / last
Qdrant
qdrant.host	127.0.0.1	Qdrant bind/connect host
qdrant.port	6333	Qdrant HTTP port
qdrant.grpc_port	6334	Qdrant gRPC port
qdrant.version	v1.17.1	Pinned Qdrant release for auto-download
qdrant.no_spawn	false	Skip auto-spawn (BYO server)
Chunker
chunker.overlap	400	Sliding overlap chars between chunks
chunker.size_code	4000	Soft target chunk size for code
chunker.size_prose	6000	Soft target chunk size for prose
chunker.hard_max	8000	Absolute max chunk size
chunker.semantic_threshold	0.5	Topic-shift threshold for SemanticChunker (lower = sharper splits)
Tagger
tagger.zero_shot	true	Enable zero-shot topic tagging
tagger.llm	false	Enable LLM topic tagging (replaces zero-shot during ingest/refresh)
tagger.llm_provider	anthropic	LLM provider: anthropic / openai / ollama / vllm / gemini / local
tagger.llm_model	claude-haiku-4-5	LLM tagger model name
tagger.llm_base_url	—	LLM tagger API base URL override
Tagger — local model (llama-cpp, provider=local)
tagger.local.model_repo	unsloth/Qwen3-1.7B-GGUF	HF repo for GGUF auto-download
tagger.local.model_file	Qwen3-1.7B-Q4_K_M.gguf	GGUF filename within the repo
tagger.local.model_path	—	Absolute path to a local GGUF (overrides repo/file)
tagger.local.n_ctx	8192	Tagger context window in tokens
tagger.local.n_gpu_layers	-1	GPU layers to offload (-1 = all)
Ingest (docs + URLs)
ingest.doc_formats	pdf,docx,pptx,xlsx,csv,epub,rtf,html,eml,json	Comma-separated doc formats enabled for the file walker
ingest.ocr_enabled	false	OCR for scanned PDFs + images (requires tesseract)
ingest.ocr_lang	eng	Tesseract language codes (e.g. eng+fra)
ingest.max_doc_size_mb	25	Per-file byte cap for document extraction
ingest.url_timeout_sec	30	HTTP connect timeout for URL fetch
ingest.url_read_timeout_sec	300	Per-chunk read timeout for URL fetch (raise for very large files)
ingest.url_user_agent	imprint/1.0	HTTP User-Agent header
ingest.url_respect_robots	true	Check robots.txt before fetching
Chat (dashboard panel)
chat.enabled	true	Enable the dashboard chat panel
chat.provider	local	Chat provider: local / vllm / openai / ollama / gemini / anthropic
chat.model	—	Model name for remote providers (default per provider)
chat.base_url	—	Base URL override for OpenAI-compat providers
chat.model_repo	unsloth/gemma-4-E4B-it-GGUF	HF repo for GGUF auto-download (local provider)
chat.model_file	gemma-4-E4B-it-Q4_K_M.gguf	GGUF filename within the repo
chat.model_path	—	Absolute path to a local GGUF (overrides repo/file)
chat.n_ctx	16384	Chat context window tokens
chat.n_gpu_layers	-1	GPU layers to offload (-1 = all)
chat.max_tokens	1024	Max tokens per chat response
chat.temperature	0.3	Chat sampling temperature
chat.max_tool_iters	6	Max tool-call iterations per chat turn
Session summarizer (opt-in, Stop hook)
summarizer.enabled	false	Enable the LLM-based session summarizer
summarizer.provider	ollama	Provider: ollama / vllm / anthropic / openai / gemini
summarizer.model	qwen3:1.7b	Model name for the chosen provider
summarizer.base_url	—	API base URL override
summarizer.min_messages	5	Skip sessions with fewer messages
summarizer.max_input_tokens	20000	Truncate transcript before summarizing
Other
collection	memories	Default Qdrant collection name (workspace suffix is appended automatically)

Each setting also has an IMPRINT_* env var (shown via imprint config get <key>). Env vars still work and take priority over config.json.

Env-only settings

These don’t live in config.json because they’re path-shaped infra knobs the config system is not designed to manage. Set them at process start:

Env var	Default	Purpose
IMPRINT_DATA_DIR	<repo>/data or ~/.local/share/imprint/data	Root for workspaces, Qdrant storage, SQLite graphs, config.json. Override for multi-install setups or when running out of a read-only tree.
IMPRINT_QDRANT_BIN	auto (downloaded into data/qdrant-bin/)	Path to a system-installed qdrant binary, bypassing the pinned auto-download.
IMPRINT_MCP_IDLE_S	30	Seconds of MCP inactivity after which the server releases the embedded Qdrant client lock so a separate imprint ingest can grab it.