my-project-public

repository

loading code, commits, and activity

repositories

loading repo index

my-project-public — gitlawb

#1	# Clawd Memory Installation Guide For LLM Agents
#2
#3	> Target audience: AI agents that need to install, configure, and verify Clawd Memory for a user.
#4	> Use this when the user asks to set up Clawd memory, agent memory, Hermes memory, or local persistent memory.
#5
#6	## Decision Path
#7
#8	\| User has \| Use \|
#9	\|---\|---\|
#10	\| This source checkout \| Run `python3 -m mnemosyne.clawd_brain ...` directly \|
#11	\| Wants editable development install \| `python3 -m pip install -e ".[all,dev]"` \|
#12	\| Wants package install \| `pip install mnemosyne-memory[all]` \|
#13	\| Uses Hermes \| Register the existing Mnemosyne-compatible provider \|
#14
#15	The package and module names remain `mnemosyne-memory` and `mnemosyne` for compatibility. The Clawd-facing CLI is `clawd-brain`.
#16
#17	## Source Checkout Setup
#18
#19	```bash
#20	cd /Users/8bit/bots/Cladwbot-solana/solana-clawd/MemeBRain
#21	python3 -m mnemosyne.clawd_brain init
#22	python3 -m mnemosyne.clawd_brain status
#23	```
#24
#25	If imports fail, install editable:
#26
#27	```bash
#28	python3 -m pip install -e ".[all,dev]"
#29	```
#30
#31	## Verify Clawd Memory
#32
#33	Run a write and recall:
#34
#35	```bash
#36	python3 -m mnemosyne.clawd_brain remember \
#37	"Clawd Memory Install Check" \
#38	"Clawd Memory was initialized and verified locally." \
#39	--kind agent \
#40	--tag clawd \
#41	--importance 0.8
#42
#43	python3 -m mnemosyne.clawd_brain recall "install check" --top-k 3
#44	```
#45
#46	Expected result: JSON containing at least one matching memory or vault note.
#47
#48	## Hermes Setup
#49
#50	Install or register the provider:
#51
#52	```bash
#53	pip install mnemosyne-memory[all]
#54	python -m mnemosyne.install
#55	```
#56
#57	Configure Hermes:
#58
#59	```bash
#60	hermes config set memory.provider mnemosyne
#61	```
#62
#63	Or edit `~/.hermes/config.yaml`:
#64
#65	```yaml
#66	memory:
#67	provider: mnemosyne
#68	plugins:
#69	enabled:
#70	- mnemosyne
#71	```
#72
#73	Verify:
#74
#75	```bash
#76	hermes gateway restart
#77	hermes memory status
#78	hermes mnemosyne stats
#79	hermes tools list \| grep mnemosyne
#80	```
#81
#82	Expected provider/tool names still contain `mnemosyne`. Treat those as compatibility names for Clawd Memory.
#83
#84	## Agent Usage Rules
#85
#86	- Always recall before answering if the task depends on prior Clawd context.
#87	- Use `remember` for durable facts, preferences, decisions, and risk findings.
#88	- Use `research` for URLs and investigation topics.
#89	- Do not store secrets or private keys.
#90	- Prefer `kind`, `source`, `tag`, and `importance` on every durable memory.
#91
#92	## Environment
#93
#94	\| Variable \| Use \|
#95	\|---\|---\|
#96	\| `CLAWD_BRAIN_VAULT` \| Override the Clawd markdown vault \|
#97	\| `MNEMOSYNE_DATA_DIR` \| Override the SQLite engine data directory \|
#98	\| `MNEMOSYNE_HOST_LLM_ENABLED` \| Route consolidation through a host framework LLM \|
#99	\| `MNEMOSYNE_LLM_BASE_URL` \| Use an OpenAI-compatible consolidation endpoint \|
#100	\| `MNEMOSYNE_LLM_API_KEY` \| API key for that endpoint \|
#101	\| `MNEMOSYNE_LLM_MODEL` \| Consolidation model \|
#102
#103	## Troubleshooting
#104
#105	### `No module named mnemosyne`
#106
#107	Install from the source directory:
#108
#109	```bash
#110	python3 -m pip install -e ".[all]"
#111	```
#112
#113	### Hermes provider not found
#114
#115	```bash
#116	python -m mnemosyne.install
#117	hermes gateway restart
#118	hermes memory status
#119	```
#120
#121	### Recall returns nothing
#122
#123	Write a test memory, confirm `status`, and check that the same `--bank` and `CLAWD_BRAIN_VAULT` are being used.
#124

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public