ai-agent-portable-memory 0.1.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- ai_agent_portable_memory-0.1.0/PKG-INFO +281 -0
- ai_agent_portable_memory-0.1.0/README.md +257 -0
- ai_agent_portable_memory-0.1.0/ai_agent_portable_memory.egg-info/PKG-INFO +281 -0
- ai_agent_portable_memory-0.1.0/ai_agent_portable_memory.egg-info/SOURCES.txt +19 -0
- ai_agent_portable_memory-0.1.0/ai_agent_portable_memory.egg-info/dependency_links.txt +1 -0
- ai_agent_portable_memory-0.1.0/ai_agent_portable_memory.egg-info/entry_points.txt +2 -0
- ai_agent_portable_memory-0.1.0/ai_agent_portable_memory.egg-info/requires.txt +19 -0
- ai_agent_portable_memory-0.1.0/ai_agent_portable_memory.egg-info/top_level.txt +1 -0
- ai_agent_portable_memory-0.1.0/memory/__init__.py +1 -0
- ai_agent_portable_memory-0.1.0/memory/backends/__init__.py +0 -0
- ai_agent_portable_memory-0.1.0/memory/backends/base.py +41 -0
- ai_agent_portable_memory-0.1.0/memory/backends/firebase.py +82 -0
- ai_agent_portable_memory-0.1.0/memory/backends/s3.py +148 -0
- ai_agent_portable_memory-0.1.0/memory/backends/sqlite.py +128 -0
- ai_agent_portable_memory-0.1.0/memory/backends/supabase.py +83 -0
- ai_agent_portable_memory-0.1.0/memory/cli.py +255 -0
- ai_agent_portable_memory-0.1.0/memory/compressor.py +143 -0
- ai_agent_portable_memory-0.1.0/memory/core.py +89 -0
- ai_agent_portable_memory-0.1.0/memory/editor.py +17 -0
- ai_agent_portable_memory-0.1.0/pyproject.toml +36 -0
- ai_agent_portable_memory-0.1.0/setup.cfg +4 -0
|
@@ -0,0 +1,281 @@
|
|
|
1
|
+
Metadata-Version: 2.4
|
|
2
|
+
Name: ai-agent-portable-memory
|
|
3
|
+
Version: 0.1.0
|
|
4
|
+
Summary: Portable compressed memory for AI agents — works with any LLM
|
|
5
|
+
License: MIT
|
|
6
|
+
Keywords: ai,memory,llm,chatgpt,claude,portable
|
|
7
|
+
Requires-Python: >=3.10
|
|
8
|
+
Description-Content-Type: text/markdown
|
|
9
|
+
Requires-Dist: click>=8.1
|
|
10
|
+
Requires-Dist: rich>=13.0
|
|
11
|
+
Requires-Dist: openai>=1.0
|
|
12
|
+
Requires-Dist: anthropic>=0.25
|
|
13
|
+
Requires-Dist: python-dotenv>=1.0
|
|
14
|
+
Provides-Extra: supabase
|
|
15
|
+
Requires-Dist: supabase>=2.0; extra == "supabase"
|
|
16
|
+
Provides-Extra: s3
|
|
17
|
+
Requires-Dist: boto3>=1.34; extra == "s3"
|
|
18
|
+
Provides-Extra: firebase
|
|
19
|
+
Requires-Dist: firebase-admin>=6.0; extra == "firebase"
|
|
20
|
+
Provides-Extra: all
|
|
21
|
+
Requires-Dist: supabase>=2.0; extra == "all"
|
|
22
|
+
Requires-Dist: boto3>=1.34; extra == "all"
|
|
23
|
+
Requires-Dist: firebase-admin>=6.0; extra == "all"
|
|
24
|
+
|
|
25
|
+
# AI Agent Portable Memory
|
|
26
|
+
|
|
27
|
+
> Carry your context across any AI — ChatGPT, Claude, Gemini, Grok, or any LLM.
|
|
28
|
+
|
|
29
|
+
Instead of losing your work every time you switch AI tools or start a new chat, **AI Agent Portable Memory** lets you save compact, structured session summaries and load them instantly into any AI conversation.
|
|
30
|
+
|
|
31
|
+
---
|
|
32
|
+
|
|
33
|
+
## The Problem
|
|
34
|
+
|
|
35
|
+
Every AI session starts blank. You lose:
|
|
36
|
+
- What you were building
|
|
37
|
+
- Decisions you already made
|
|
38
|
+
- Where you stopped
|
|
39
|
+
|
|
40
|
+
Copy-pasting raw chat logs wastes too many context tokens and is unmanageable.
|
|
41
|
+
|
|
42
|
+
## The Solution
|
|
43
|
+
|
|
44
|
+
A portable memory store that:
|
|
45
|
+
1. **Compresses** your session into a structured, token-efficient summary using an LLM
|
|
46
|
+
2. **Lets you review and edit** the summary before saving
|
|
47
|
+
3. **Stores it** locally (SQLite) or in the cloud (Supabase, S3/R2, Firebase)
|
|
48
|
+
4. **Loads it** as a clean context block you paste into any AI
|
|
49
|
+
|
|
50
|
+
---
|
|
51
|
+
|
|
52
|
+
## Install
|
|
53
|
+
|
|
54
|
+
```bash
|
|
55
|
+
pip install ai-agent-portable-memory
|
|
56
|
+
```
|
|
57
|
+
|
|
58
|
+
Or from source:
|
|
59
|
+
|
|
60
|
+
```bash
|
|
61
|
+
git clone https://github.com/your-username/AI-Agent-Portable-Memory
|
|
62
|
+
cd AI-Agent-Portable-Memory
|
|
63
|
+
pip install -e .
|
|
64
|
+
```
|
|
65
|
+
|
|
66
|
+
For cloud storage backends:
|
|
67
|
+
|
|
68
|
+
```bash
|
|
69
|
+
pip install "ai-agent-portable-memory[supabase]" # Supabase
|
|
70
|
+
pip install "ai-agent-portable-memory[s3]" # AWS S3 or Cloudflare R2
|
|
71
|
+
pip install "ai-agent-portable-memory[firebase]" # Firebase Firestore
|
|
72
|
+
pip install "ai-agent-portable-memory[all]" # Everything
|
|
73
|
+
```
|
|
74
|
+
|
|
75
|
+
---
|
|
76
|
+
|
|
77
|
+
## Quick Start
|
|
78
|
+
|
|
79
|
+
### 1. Set your LLM API key
|
|
80
|
+
|
|
81
|
+
```bash
|
|
82
|
+
export OPENAI_API_KEY=sk-... # OpenAI (default)
|
|
83
|
+
# or
|
|
84
|
+
export ANTHROPIC_API_KEY=sk-ant-... # Claude
|
|
85
|
+
# or
|
|
86
|
+
export GROQ_API_KEY=gsk_... # Groq (free tier available)
|
|
87
|
+
```
|
|
88
|
+
|
|
89
|
+
### 2. Save a session
|
|
90
|
+
|
|
91
|
+
```bash
|
|
92
|
+
memory save
|
|
93
|
+
```
|
|
94
|
+
|
|
95
|
+
Paste your chat transcript, press `Ctrl+D`. The LLM compresses it into a structured summary, your editor opens for review, and it's saved.
|
|
96
|
+
|
|
97
|
+
### 3. List your sessions
|
|
98
|
+
|
|
99
|
+
```bash
|
|
100
|
+
memory list
|
|
101
|
+
```
|
|
102
|
+
|
|
103
|
+
```
|
|
104
|
+
╭────┬──────────┬──────────────────────────────────┬───────────┬──────────────────╮
|
|
105
|
+
│ # │ ID │ Title │ Project │ Updated │
|
|
106
|
+
├────┼──────────┼──────────────────────────────────┼───────────┼──────────────────┤
|
|
107
|
+
│ 1 │ a3f2c1b0 │ JWT Auth Middleware — halfway │ ecommerce │ 2026-11-18 14:30 │
|
|
108
|
+
│ 2 │ 9e1d4a72 │ PostgreSQL schema design │ ecommerce │ 2026-11-17 09:15 │
|
|
109
|
+
│ 3 │ 7c8b3f91 │ React dashboard setup │ dashboard │ 2026-11-16 18:00 │
|
|
110
|
+
╰────┴──────────┴──────────────────────────────────┴───────────┴──────────────────╯
|
|
111
|
+
```
|
|
112
|
+
|
|
113
|
+
### 4. Load context into any AI
|
|
114
|
+
|
|
115
|
+
```bash
|
|
116
|
+
memory load 1 # by list number
|
|
117
|
+
memory load a3f2c1b0 # by ID prefix
|
|
118
|
+
memory load 1 --copy # copy to clipboard
|
|
119
|
+
```
|
|
120
|
+
|
|
121
|
+
Paste the output at the start of your new AI chat. The AI instantly knows where you left off.
|
|
122
|
+
|
|
123
|
+
---
|
|
124
|
+
|
|
125
|
+
## What a session summary looks like
|
|
126
|
+
|
|
127
|
+
```markdown
|
|
128
|
+
## Title
|
|
129
|
+
JWT Auth Middleware — halfway through refresh token rotation
|
|
130
|
+
|
|
131
|
+
## Project
|
|
132
|
+
ecommerce
|
|
133
|
+
|
|
134
|
+
## What we were doing
|
|
135
|
+
Building JWT authentication for a React + Node.js e-commerce app.
|
|
136
|
+
|
|
137
|
+
## What was completed
|
|
138
|
+
- Login page UI (src/pages/Login.tsx)
|
|
139
|
+
- JWT token generation on backend (src/auth/jwt.js)
|
|
140
|
+
- httpOnly cookie setup on login endpoint
|
|
141
|
+
|
|
142
|
+
## Key decisions made
|
|
143
|
+
- httpOnly cookies instead of localStorage — XSS protection
|
|
144
|
+
- Access token TTL: 15min, Refresh token TTL: 7 days
|
|
145
|
+
- Library: jsonwebtoken v9
|
|
146
|
+
|
|
147
|
+
## Current state / where we stopped
|
|
148
|
+
Halfway through refresh token rotation middleware.
|
|
149
|
+
File: src/middleware/refreshToken.js — function written, error handling missing.
|
|
150
|
+
|
|
151
|
+
## Next steps
|
|
152
|
+
1. Finish error handling in refreshToken.js
|
|
153
|
+
2. Add protected route wrapper component
|
|
154
|
+
3. Test full auth flow with Postman
|
|
155
|
+
|
|
156
|
+
## Important context
|
|
157
|
+
- Backend: Express.js on port 3001
|
|
158
|
+
- Frontend: React + Vite on port 5173
|
|
159
|
+
- DB: PostgreSQL with Prisma ORM
|
|
160
|
+
- Auth route prefix: /api/auth
|
|
161
|
+
|
|
162
|
+
## Tags
|
|
163
|
+
react, nodejs, jwt, auth, express, postgresql
|
|
164
|
+
```
|
|
165
|
+
|
|
166
|
+
**Compact. Structured. Ready to paste.**
|
|
167
|
+
|
|
168
|
+
---
|
|
169
|
+
|
|
170
|
+
## Configuration
|
|
171
|
+
|
|
172
|
+
Create a `.env` file or set environment variables:
|
|
173
|
+
|
|
174
|
+
```env
|
|
175
|
+
# Storage backend (default: sqlite)
|
|
176
|
+
MEMORY_BACKEND=sqlite
|
|
177
|
+
|
|
178
|
+
# LLM provider for compression (default: openai)
|
|
179
|
+
MEMORY_LLM_PROVIDER=openai
|
|
180
|
+
|
|
181
|
+
# LLM API keys
|
|
182
|
+
OPENAI_API_KEY=sk-...
|
|
183
|
+
ANTHROPIC_API_KEY=sk-ant-...
|
|
184
|
+
GROQ_API_KEY=gsk_...
|
|
185
|
+
```
|
|
186
|
+
|
|
187
|
+
### Storage Backends
|
|
188
|
+
|
|
189
|
+
#### SQLite (default — no setup required)
|
|
190
|
+
|
|
191
|
+
```env
|
|
192
|
+
MEMORY_BACKEND=sqlite
|
|
193
|
+
SQLITE_DB_PATH=~/.ai-memory/memory.db # optional, this is the default
|
|
194
|
+
```
|
|
195
|
+
|
|
196
|
+
#### Supabase
|
|
197
|
+
|
|
198
|
+
```env
|
|
199
|
+
MEMORY_BACKEND=supabase
|
|
200
|
+
SUPABASE_URL=https://xxxxx.supabase.co
|
|
201
|
+
SUPABASE_KEY=your-anon-key
|
|
202
|
+
```
|
|
203
|
+
|
|
204
|
+
Run this SQL in your Supabase project:
|
|
205
|
+
|
|
206
|
+
```sql
|
|
207
|
+
CREATE TABLE sessions (
|
|
208
|
+
id TEXT PRIMARY KEY,
|
|
209
|
+
project TEXT NOT NULL DEFAULT 'default',
|
|
210
|
+
title TEXT NOT NULL,
|
|
211
|
+
summary TEXT NOT NULL,
|
|
212
|
+
tags JSONB NOT NULL DEFAULT '[]',
|
|
213
|
+
created_at TIMESTAMPTZ NOT NULL DEFAULT now(),
|
|
214
|
+
updated_at TIMESTAMPTZ NOT NULL DEFAULT now()
|
|
215
|
+
);
|
|
216
|
+
CREATE INDEX ON sessions USING GIN (to_tsvector('english', title || ' ' || summary));
|
|
217
|
+
```
|
|
218
|
+
|
|
219
|
+
#### Cloudflare R2 / AWS S3
|
|
220
|
+
|
|
221
|
+
```env
|
|
222
|
+
MEMORY_BACKEND=s3
|
|
223
|
+
S3_BUCKET_NAME=my-memory-bucket
|
|
224
|
+
AWS_ACCESS_KEY_ID=...
|
|
225
|
+
AWS_SECRET_ACCESS_KEY=...
|
|
226
|
+
|
|
227
|
+
# For Cloudflare R2 (add this):
|
|
228
|
+
S3_ENDPOINT_URL=https://<account_id>.r2.cloudflarestorage.com
|
|
229
|
+
```
|
|
230
|
+
|
|
231
|
+
#### Firebase Firestore
|
|
232
|
+
|
|
233
|
+
```env
|
|
234
|
+
MEMORY_BACKEND=firebase
|
|
235
|
+
FIREBASE_CREDENTIALS_PATH=./serviceAccountKey.json
|
|
236
|
+
FIREBASE_PROJECT_ID=my-firebase-project
|
|
237
|
+
```
|
|
238
|
+
|
|
239
|
+
---
|
|
240
|
+
|
|
241
|
+
## CLI Reference
|
|
242
|
+
|
|
243
|
+
```
|
|
244
|
+
memory list List all sessions
|
|
245
|
+
memory list --project myapp Filter by project
|
|
246
|
+
|
|
247
|
+
memory load <id> Print session context
|
|
248
|
+
memory load <id> --copy Copy context to clipboard
|
|
249
|
+
|
|
250
|
+
memory save Save new session (paste transcript)
|
|
251
|
+
memory save --provider anthropic Use a specific LLM provider
|
|
252
|
+
memory save --project myapp Set project name
|
|
253
|
+
memory save --no-edit Skip editor review
|
|
254
|
+
|
|
255
|
+
memory search <query> Search sessions by keyword
|
|
256
|
+
memory search auth --project api Filter search by project
|
|
257
|
+
|
|
258
|
+
memory delete <id> Delete a session
|
|
259
|
+
|
|
260
|
+
memory config Show current configuration
|
|
261
|
+
```
|
|
262
|
+
|
|
263
|
+
---
|
|
264
|
+
|
|
265
|
+
## How it works
|
|
266
|
+
|
|
267
|
+
```
|
|
268
|
+
Your chat → memory save → LLM compresses → You review → Stored
|
|
269
|
+
↓
|
|
270
|
+
New AI chat ← paste ← memory load ← compact summary ← Retrieved
|
|
271
|
+
```
|
|
272
|
+
|
|
273
|
+
---
|
|
274
|
+
|
|
275
|
+
## Contributing
|
|
276
|
+
|
|
277
|
+
PRs welcome. To add a new storage backend, implement the `BaseBackend` interface in `memory/backends/base.py`.
|
|
278
|
+
|
|
279
|
+
## License
|
|
280
|
+
|
|
281
|
+
MIT
|
|
@@ -0,0 +1,257 @@
|
|
|
1
|
+
# AI Agent Portable Memory
|
|
2
|
+
|
|
3
|
+
> Carry your context across any AI — ChatGPT, Claude, Gemini, Grok, or any LLM.
|
|
4
|
+
|
|
5
|
+
Instead of losing your work every time you switch AI tools or start a new chat, **AI Agent Portable Memory** lets you save compact, structured session summaries and load them instantly into any AI conversation.
|
|
6
|
+
|
|
7
|
+
---
|
|
8
|
+
|
|
9
|
+
## The Problem
|
|
10
|
+
|
|
11
|
+
Every AI session starts blank. You lose:
|
|
12
|
+
- What you were building
|
|
13
|
+
- Decisions you already made
|
|
14
|
+
- Where you stopped
|
|
15
|
+
|
|
16
|
+
Copy-pasting raw chat logs wastes too many context tokens and is unmanageable.
|
|
17
|
+
|
|
18
|
+
## The Solution
|
|
19
|
+
|
|
20
|
+
A portable memory store that:
|
|
21
|
+
1. **Compresses** your session into a structured, token-efficient summary using an LLM
|
|
22
|
+
2. **Lets you review and edit** the summary before saving
|
|
23
|
+
3. **Stores it** locally (SQLite) or in the cloud (Supabase, S3/R2, Firebase)
|
|
24
|
+
4. **Loads it** as a clean context block you paste into any AI
|
|
25
|
+
|
|
26
|
+
---
|
|
27
|
+
|
|
28
|
+
## Install
|
|
29
|
+
|
|
30
|
+
```bash
|
|
31
|
+
pip install ai-agent-portable-memory
|
|
32
|
+
```
|
|
33
|
+
|
|
34
|
+
Or from source:
|
|
35
|
+
|
|
36
|
+
```bash
|
|
37
|
+
git clone https://github.com/your-username/AI-Agent-Portable-Memory
|
|
38
|
+
cd AI-Agent-Portable-Memory
|
|
39
|
+
pip install -e .
|
|
40
|
+
```
|
|
41
|
+
|
|
42
|
+
For cloud storage backends:
|
|
43
|
+
|
|
44
|
+
```bash
|
|
45
|
+
pip install "ai-agent-portable-memory[supabase]" # Supabase
|
|
46
|
+
pip install "ai-agent-portable-memory[s3]" # AWS S3 or Cloudflare R2
|
|
47
|
+
pip install "ai-agent-portable-memory[firebase]" # Firebase Firestore
|
|
48
|
+
pip install "ai-agent-portable-memory[all]" # Everything
|
|
49
|
+
```
|
|
50
|
+
|
|
51
|
+
---
|
|
52
|
+
|
|
53
|
+
## Quick Start
|
|
54
|
+
|
|
55
|
+
### 1. Set your LLM API key
|
|
56
|
+
|
|
57
|
+
```bash
|
|
58
|
+
export OPENAI_API_KEY=sk-... # OpenAI (default)
|
|
59
|
+
# or
|
|
60
|
+
export ANTHROPIC_API_KEY=sk-ant-... # Claude
|
|
61
|
+
# or
|
|
62
|
+
export GROQ_API_KEY=gsk_... # Groq (free tier available)
|
|
63
|
+
```
|
|
64
|
+
|
|
65
|
+
### 2. Save a session
|
|
66
|
+
|
|
67
|
+
```bash
|
|
68
|
+
memory save
|
|
69
|
+
```
|
|
70
|
+
|
|
71
|
+
Paste your chat transcript, press `Ctrl+D`. The LLM compresses it into a structured summary, your editor opens for review, and it's saved.
|
|
72
|
+
|
|
73
|
+
### 3. List your sessions
|
|
74
|
+
|
|
75
|
+
```bash
|
|
76
|
+
memory list
|
|
77
|
+
```
|
|
78
|
+
|
|
79
|
+
```
|
|
80
|
+
╭────┬──────────┬──────────────────────────────────┬───────────┬──────────────────╮
|
|
81
|
+
│ # │ ID │ Title │ Project │ Updated │
|
|
82
|
+
├────┼──────────┼──────────────────────────────────┼───────────┼──────────────────┤
|
|
83
|
+
│ 1 │ a3f2c1b0 │ JWT Auth Middleware — halfway │ ecommerce │ 2026-11-18 14:30 │
|
|
84
|
+
│ 2 │ 9e1d4a72 │ PostgreSQL schema design │ ecommerce │ 2026-11-17 09:15 │
|
|
85
|
+
│ 3 │ 7c8b3f91 │ React dashboard setup │ dashboard │ 2026-11-16 18:00 │
|
|
86
|
+
╰────┴──────────┴──────────────────────────────────┴───────────┴──────────────────╯
|
|
87
|
+
```
|
|
88
|
+
|
|
89
|
+
### 4. Load context into any AI
|
|
90
|
+
|
|
91
|
+
```bash
|
|
92
|
+
memory load 1 # by list number
|
|
93
|
+
memory load a3f2c1b0 # by ID prefix
|
|
94
|
+
memory load 1 --copy # copy to clipboard
|
|
95
|
+
```
|
|
96
|
+
|
|
97
|
+
Paste the output at the start of your new AI chat. The AI instantly knows where you left off.
|
|
98
|
+
|
|
99
|
+
---
|
|
100
|
+
|
|
101
|
+
## What a session summary looks like
|
|
102
|
+
|
|
103
|
+
```markdown
|
|
104
|
+
## Title
|
|
105
|
+
JWT Auth Middleware — halfway through refresh token rotation
|
|
106
|
+
|
|
107
|
+
## Project
|
|
108
|
+
ecommerce
|
|
109
|
+
|
|
110
|
+
## What we were doing
|
|
111
|
+
Building JWT authentication for a React + Node.js e-commerce app.
|
|
112
|
+
|
|
113
|
+
## What was completed
|
|
114
|
+
- Login page UI (src/pages/Login.tsx)
|
|
115
|
+
- JWT token generation on backend (src/auth/jwt.js)
|
|
116
|
+
- httpOnly cookie setup on login endpoint
|
|
117
|
+
|
|
118
|
+
## Key decisions made
|
|
119
|
+
- httpOnly cookies instead of localStorage — XSS protection
|
|
120
|
+
- Access token TTL: 15min, Refresh token TTL: 7 days
|
|
121
|
+
- Library: jsonwebtoken v9
|
|
122
|
+
|
|
123
|
+
## Current state / where we stopped
|
|
124
|
+
Halfway through refresh token rotation middleware.
|
|
125
|
+
File: src/middleware/refreshToken.js — function written, error handling missing.
|
|
126
|
+
|
|
127
|
+
## Next steps
|
|
128
|
+
1. Finish error handling in refreshToken.js
|
|
129
|
+
2. Add protected route wrapper component
|
|
130
|
+
3. Test full auth flow with Postman
|
|
131
|
+
|
|
132
|
+
## Important context
|
|
133
|
+
- Backend: Express.js on port 3001
|
|
134
|
+
- Frontend: React + Vite on port 5173
|
|
135
|
+
- DB: PostgreSQL with Prisma ORM
|
|
136
|
+
- Auth route prefix: /api/auth
|
|
137
|
+
|
|
138
|
+
## Tags
|
|
139
|
+
react, nodejs, jwt, auth, express, postgresql
|
|
140
|
+
```
|
|
141
|
+
|
|
142
|
+
**Compact. Structured. Ready to paste.**
|
|
143
|
+
|
|
144
|
+
---
|
|
145
|
+
|
|
146
|
+
## Configuration
|
|
147
|
+
|
|
148
|
+
Create a `.env` file or set environment variables:
|
|
149
|
+
|
|
150
|
+
```env
|
|
151
|
+
# Storage backend (default: sqlite)
|
|
152
|
+
MEMORY_BACKEND=sqlite
|
|
153
|
+
|
|
154
|
+
# LLM provider for compression (default: openai)
|
|
155
|
+
MEMORY_LLM_PROVIDER=openai
|
|
156
|
+
|
|
157
|
+
# LLM API keys
|
|
158
|
+
OPENAI_API_KEY=sk-...
|
|
159
|
+
ANTHROPIC_API_KEY=sk-ant-...
|
|
160
|
+
GROQ_API_KEY=gsk_...
|
|
161
|
+
```
|
|
162
|
+
|
|
163
|
+
### Storage Backends
|
|
164
|
+
|
|
165
|
+
#### SQLite (default — no setup required)
|
|
166
|
+
|
|
167
|
+
```env
|
|
168
|
+
MEMORY_BACKEND=sqlite
|
|
169
|
+
SQLITE_DB_PATH=~/.ai-memory/memory.db # optional, this is the default
|
|
170
|
+
```
|
|
171
|
+
|
|
172
|
+
#### Supabase
|
|
173
|
+
|
|
174
|
+
```env
|
|
175
|
+
MEMORY_BACKEND=supabase
|
|
176
|
+
SUPABASE_URL=https://xxxxx.supabase.co
|
|
177
|
+
SUPABASE_KEY=your-anon-key
|
|
178
|
+
```
|
|
179
|
+
|
|
180
|
+
Run this SQL in your Supabase project:
|
|
181
|
+
|
|
182
|
+
```sql
|
|
183
|
+
CREATE TABLE sessions (
|
|
184
|
+
id TEXT PRIMARY KEY,
|
|
185
|
+
project TEXT NOT NULL DEFAULT 'default',
|
|
186
|
+
title TEXT NOT NULL,
|
|
187
|
+
summary TEXT NOT NULL,
|
|
188
|
+
tags JSONB NOT NULL DEFAULT '[]',
|
|
189
|
+
created_at TIMESTAMPTZ NOT NULL DEFAULT now(),
|
|
190
|
+
updated_at TIMESTAMPTZ NOT NULL DEFAULT now()
|
|
191
|
+
);
|
|
192
|
+
CREATE INDEX ON sessions USING GIN (to_tsvector('english', title || ' ' || summary));
|
|
193
|
+
```
|
|
194
|
+
|
|
195
|
+
#### Cloudflare R2 / AWS S3
|
|
196
|
+
|
|
197
|
+
```env
|
|
198
|
+
MEMORY_BACKEND=s3
|
|
199
|
+
S3_BUCKET_NAME=my-memory-bucket
|
|
200
|
+
AWS_ACCESS_KEY_ID=...
|
|
201
|
+
AWS_SECRET_ACCESS_KEY=...
|
|
202
|
+
|
|
203
|
+
# For Cloudflare R2 (add this):
|
|
204
|
+
S3_ENDPOINT_URL=https://<account_id>.r2.cloudflarestorage.com
|
|
205
|
+
```
|
|
206
|
+
|
|
207
|
+
#### Firebase Firestore
|
|
208
|
+
|
|
209
|
+
```env
|
|
210
|
+
MEMORY_BACKEND=firebase
|
|
211
|
+
FIREBASE_CREDENTIALS_PATH=./serviceAccountKey.json
|
|
212
|
+
FIREBASE_PROJECT_ID=my-firebase-project
|
|
213
|
+
```
|
|
214
|
+
|
|
215
|
+
---
|
|
216
|
+
|
|
217
|
+
## CLI Reference
|
|
218
|
+
|
|
219
|
+
```
|
|
220
|
+
memory list List all sessions
|
|
221
|
+
memory list --project myapp Filter by project
|
|
222
|
+
|
|
223
|
+
memory load <id> Print session context
|
|
224
|
+
memory load <id> --copy Copy context to clipboard
|
|
225
|
+
|
|
226
|
+
memory save Save new session (paste transcript)
|
|
227
|
+
memory save --provider anthropic Use a specific LLM provider
|
|
228
|
+
memory save --project myapp Set project name
|
|
229
|
+
memory save --no-edit Skip editor review
|
|
230
|
+
|
|
231
|
+
memory search <query> Search sessions by keyword
|
|
232
|
+
memory search auth --project api Filter search by project
|
|
233
|
+
|
|
234
|
+
memory delete <id> Delete a session
|
|
235
|
+
|
|
236
|
+
memory config Show current configuration
|
|
237
|
+
```
|
|
238
|
+
|
|
239
|
+
---
|
|
240
|
+
|
|
241
|
+
## How it works
|
|
242
|
+
|
|
243
|
+
```
|
|
244
|
+
Your chat → memory save → LLM compresses → You review → Stored
|
|
245
|
+
↓
|
|
246
|
+
New AI chat ← paste ← memory load ← compact summary ← Retrieved
|
|
247
|
+
```
|
|
248
|
+
|
|
249
|
+
---
|
|
250
|
+
|
|
251
|
+
## Contributing
|
|
252
|
+
|
|
253
|
+
PRs welcome. To add a new storage backend, implement the `BaseBackend` interface in `memory/backends/base.py`.
|
|
254
|
+
|
|
255
|
+
## License
|
|
256
|
+
|
|
257
|
+
MIT
|