Skip to content

cli.md

Latest commit

 

History

History
358 lines (308 loc) · 8.94 KB

cli.md

File metadata and controls

358 lines (308 loc) · 8.94 KB

CLI Reference

memory.cpp ships a practical command line for one job: helping your repo remember what happened, why it changed, and what to do next.

Beginner-friendly commands

These are the first commands to try.

memory welcome
memory setup --interactive
memory what
memory where
memory today
memory yesterday
memory next
memory show-map
memory show-context
memory show-inbox
memory privacy status

Core commands

memory init
memory remember
memory recall
memory explain
memory forget
memory patch
memory context
memory compile
memory import
memory watch
memory timeline
memory graph
memory stats
memory list
memory export
memory workspace
memory policy
memory snapshot
memory diff
memory inbox
memory attach
memory serve
memory dashboard
memory proxy
memory mcp

Launch-polish commands

These are routed through a small pre-parser so the launch build avoids a known Clap stack-overflow edge case from an oversized nested command tree.

memory edit
memory restore
memory demo
memory audit-log
memory doctor
memory dev
memory extract
memory git
memory ignore
memory map
memory start
memory stop
memory status
memory setup
memory tutorial
memory terminal
memory ci
memory embeddings
memory privacy
memory pack
memory token-firewall
memory firewall
memory cache-plan
memory kv-report
memory kv-compile
memory prefill-report
memory kv-budget
memory signal-density
memory batch-plan
memory runtime-profile
memory cache-audit
memory trace-rollup
memory trace
memory mistake
memory mistakes
memory conflicts
memory stale
memory resolve
memory savings
memory runtime-plan
memory bench-context
memory bench
memory explain-compile
memory roi
memory leaderboard
memory cache-hash
memory cache-stability
memory memories
memory update-memory
memory profile
memory entity
memory session
memory insight
memory trust-report
memory redactions
memory evidence
memory quarantine
memory review
memory flight
memory context-diff
memory blame
memory explain-pack
memory test
memory ci-check
memory ask
memory suggest
memory warnings
memory proactive
memory ingest
memory shared-context
memory heatmap
memory report
memory agents-score
memory badge
memory recipe
memory preflight
memory demo multi-model
memory docs list
memory docs summarize
memory docs search
memory examples list
memory examples run <name>

Most important workflows

Friendly first run

memory welcome
memory setup --developer
memory what
memory where
memory privacy status

Remember and recall

memory remember "Use SQLite for local-first durability." --workspace demo --kind decision
memory recall "why SQLite" --workspace demo --content
memory explain "why SQLite" --workspace demo

Daily development flow

memory dev morning --workspace demo
memory dev resume "proxy launch" --workspace demo
memory dev explain-repo . --workspace demo
memory dev next --workspace demo
memory dev context --for cursor --workspace demo

Git-aware extraction

memory git summary --since 14d
memory git decisions --since 30d
memory git bugs --since 30d
memory git ingest --workspace demo --since 14d
memory git watch --once --workspace demo

Candidate inbox

memory inbox list --workspace demo
memory inbox stats --workspace demo
memory inbox explain <id>
memory inbox edit <id> "Better wording" --kind decision
memory inbox approve-all --confidence-above 0.9

Maps

memory map . --workspace demo --type evolution --output html --save .memory.cpp/demo/evolution.html
memory map why "MCP integration" --workspace demo --output markdown
memory map impact "SQLite storage" --workspace demo --output markdown
memory show-map --workspace demo

Terminal memory

memory terminal enable
memory terminal record "cargo test" --exit-code 0 --duration-ms 12000
memory terminal last-error
memory terminal search "how did I run tests?"

Terminal memory is opt-in.

AI context packs

memory dev context --for cursor --workspace demo
memory dev context --for codex --workspace demo
memory dev context --for claude --workspace demo

Context compiler and token firewall

memory compile "fix checkout bug" --provider openai --budget 1500
memory token-firewall "fix checkout bug" --provider openai --budget 2000
memory cache-plan "answer support ticket" --provider claude
memory kv-report "summarize customer history"
memory prefill-report "fix checkout bug"
memory kv-budget "fix checkout bug" --max-kv-tokens 4096
memory signal-density "fix checkout bug"
memory batch-plan --file tests/fixtures/inference/multi_request_batch.json --provider openai
memory runtime-profile list
memory runtime-plan "fix checkout bug" --runtime llama.cpp
memory cache-audit --file tests/fixtures/inference/provider_cache_bad_order.md --provider openai
memory trace-rollup --from tests/fixtures/inference/agent_trace_long.json --every 50
memory trace compress --file examples/agent-log.txt
memory mistake "Use pnpm only. Never npm."
memory doctor "add CSV export" --provider gemini
memory pack "fix checkout bug" --for codex --budget 1500
memory pack "fix checkout bug" --for gemini --budget 1500
memory pack "fix checkout bug" --for mcp --budget 1500
memory demo multi-model
memory savings

These commands are local-first prompt hygiene tools. They compile smaller task context, report estimated prompt-side KV pressure avoided, improve signal density, group shared prefixes, and keep stale or risky memory out of AI prompts. They do not directly control provider KV caches or low-level serving kernels.

memory doctor "<task>" --provider <provider> includes an Inference Cost Stack section with raw and compiled context tokens, fresh suffix tokens, cacheable prefix tokens, omitted tokens, estimated prefill reduction, KV positions avoided, signal density, duplicate/stale/tool-trace tokens blocked, provider cache strategy, and runtime strategy.

Local context control plane

memory remember "Use cargo test before release" --scope repo --type rule
memory memories list
memory memories show <memory_id>
memory memories export --format json --output .memory.cpp/memories.json
memory profile update "Prefers concise PR summaries" --scope user
memory entity create --type agent --name Codex
memory entity create --type project --name memory.cpp
memory entity link Codex memory.cpp --relation works_on
memory session start --name release-polish --goal "finish launch readiness"
memory session add-event --type test --text "cargo test passed"
memory session summarize
memory insight derive --scope repo
memory insight list --scope repo
memory explain-compile "fix checkout bug" --provider openai
memory ask "what broke last time checkout changed?"
memory suggest "fix checkout bug"
memory warnings "change auth flow"
memory proactive --task "prepare release"
memory trust-report
memory redactions
memory evidence <memory_id>
memory quarantine review
memory review
memory flight start --goal "fix checkout bug" --tool codex
memory flight summarize
memory context-diff latest previous
memory blame <memory_id>
memory explain-pack latest
memory test
memory ci-check
memory ingest file README.md
memory shared-context export
memory heatmap --html
memory report --html
memory dashboard --html
memory agents-score --for codex
memory badge --for codex
memory recipe list
memory preflight --for codex "fix checkout bug"
memory roi --input-cost 2.50
memory leaderboard

These commands intentionally use local memory, existing metadata, generated packs, and static reports. Huge future features such as hosted dashboards, cloud sync, or kernel-level runtime optimization are not enabled by these commands.

Embeddings

memory embeddings status
memory embeddings list
memory embeddings set fastembed
memory embeddings migrate --to fastembed --dry-run

fastembed in this launch build is a zero-dependency local semantic-hashing backend. Real ONNX Runtime integration is documented as a later backend.

Runtime management

memory start --workspace demo --proxy
memory status
memory stop
memory audit-log --limit 20

Output guidance

  • Use --json for automation.
  • Use --save <path> with maps and reports you want to keep.
  • Use doctor before sharing a demo or attaching agents.
  • Create .memoryignore before importing or watching a real repository.
  • Use proxy --learn --approval-required when you want passive learning without unattended direct writes.

Virality and adoption loop

Shareable artifacts:

memory share status
memory share map --output .memory.cpp/share/project-evolution-map.html
memory share context --private-safe

Repo documentation:

memory docs generate --dry-run
memory docs generate --apply

PR and handoff workflows:

memory pr summary --base main
memory pr checklist --output .memory.cpp/share/pr-checklist.md
memory handoff new-dev --output .memory.cpp/handoff

Repo time machine and local activation:

memory timeline week
memory rewind last-week
memory changed --since "7 days ago"
memory adoption checklist
memory release-check