stash

Name: stash
Author: alash3al

byalash3al· 1 MCP server

Stars

759

Forks

How to install stash?

alash3al/stash

$git clone https://github.com/alash3al/stash

Installs into the current project.

›Prefer a prompt? Paste this to your agent

Install & use

Install stash by running `git clone https://github.com/alash3al/stash`, then use it for the current task and follow its documentation at https://github.com/alash3al/stash.

Files · 1

View on GitHub

README.md

1	# Stash
2
3	Your AI has amnesia. We fixed it.
4
5	Every LLM starts every conversation from zero. Stash gives your agent persistent memory — it remembers, recalls, consolidates, and learns across sessions. No more explaining yourself from scratch.
6
7	Open source. Self-hosted. Works with any MCP-compatible agent.
8
9	---
10
11	> Don't want to self-host?
12	> [usestash.io](https://usestash.io) is the hosted cloud version — sign in with Google, copy one MCP URL, and you're done. Free to start.
13
14	---
15
16	## Quick Start
17
18	```bash
19	git clone https://github.com/alash3al/stash.git
20	cd stash
21	cp .env.example .env # edit with your API key + model
22	docker compose up
23	```
24
25	That's it. Postgres + pgvector, migrations, MCP server with background consolidation — all in one command.
26
27	Next: [Getting Started guide](docs/GETTING_STARTED.md) — connect your MCP client, run `init` / `remember` / `recall`, and verify everything works.
28
29	Fully local (no cloud API): [Ollama setup guide](docs/LOCAL_OLLAMA.md) — host Ollama + Docker Compose, private embeddings and reasoner.
30
31	## Atlas Cloud
32
33	> [Atlas Cloud](https://www.atlascloud.ai/?utm_source=github&utm_medium=link&utm_campaign=stash) is a full-modal AI inference platform that gives developers a single AI API to access video generation, image generation, and LLM APIs. Instead of managing multiple vendor integrations, you connect once and get unified access to 300+ curated models across all modalities.
34	>
35	> Check out Atlas Cloud's coding plan promotion: [https://www.atlascloud.ai/console/coding-plan](https://www.atlascloud.ai/console/coding-plan)
36
37	Stash already supports Atlas Cloud through its OpenAI-compatible API. Set your `.env` like this:
38
39	```bash
40	STASH_OPENAI_API_KEY=your-atlas-cloud-api-key
41	STASH_OPENAI_BASE_URL=https://api.atlascloud.ai/v1
42	STASH_EMBEDDING_MODEL=text-embedding-3-small
43	STASH_REASONER_MODEL=deepseek-ai/DeepSeek-V3-0324
44	STASH_VECTOR_DIM=1536
45	```
46
47	Atlas Cloud works well here because Stash only needs:
48
49	- an embeddings model for vectorization
50	- a chat-capable reasoning model for consolidation
51	- an OpenAI-compatible base URL and API key
52
53	See [Getting Started](docs/GETTING_STARTED.md) for a fuller configuration checklist.
54
55	## MCP Client Setup
56
57	After `docker compose up`, Stash exposes an MCP server over SSE at:
58
59	```
60	http://localhost:8080/sse
61	```
62
63	Point any MCP-compatible client at that URL. Example configs:
64
65	Cursor — `~/.cursor/mcp.json`
66	```json
67	{
68	"mcpServers": {
69	"stash": {
70	"url": "http://localhost:8080/sse"
71	}
72	}
73	}
74	```
75
76	Claude Desktop — `claude_desktop_config.json`
77	```json
78	{
79	"mcpServers": {
80	"stash": {
81	"url": "http://localhost:8080/sse"
82	}
83	}
84	}
85	```
86
87	OpenCode — `~/.config/opencode/config.json`
88	```json
89	{
90	"mcp": {
91	"stash": {
92	"type": "remote",
93	"url": "http://localhost:8080/sse",
94	"enabled": true
95	}
96	}
97	}
98	```
99
100	Windsurf — `~/.codeium/windsurf/mcp_config.json`
101	```json
102	{
103	"mcpServers": {
104	"stash": {
105	"url": "http://localhost:8080/sse"
106	}
107	}
108	}
109	```
110
111	Works with any agent that supports MCP over SSE — Claude Desktop, Cursor, Windsurf, Cline, Continue, OpenAI Agents, Ollama, OpenRouter, Atlas Cloud-backed setups, and more.
112
113	## What It Does
114
115	Stash is a cognitive layer between your AI agent and the world. Episodes become facts. Facts become relationships. Relationships become patterns. Patterns become wisdom.
116
117	A 9-stage consolidation pipeline turns raw observations into structured knowledge — facts, relationships, causal links, patterns, contradictions, goal tracking, failure patterns, and hypothesis verification. Each stage only processes new data since the last run.
118
119	## Stash Cloud (Beta — Free)
120
121	A hosted, multi-tenant version of Stash is available at [usestash.io](https://usestash.io/) and is currently free while in beta.
122
123	The cloud version is written from scratch — it shares no code with this repository. It is designed from the ground up for scalability, multi-tenancy, and long-term sustainability as a product. Feature set

6 picks

Type

stash

byalash3al· 1 MCP server

Stars

759

Forks

How to install stash?

alash3al/stash

$git clone https://github.com/alash3al/stash

Installs into the current project.

›Prefer a prompt? Paste this to your agent

Install & use

Install stash by running `git clone https://github.com/alash3al/stash`, then use it for the current task and follow its documentation at https://github.com/alash3al/stash.

Files · 1

View on GitHub

README.md

1	# Stash
2
3	Your AI has amnesia. We fixed it.
4
5	Every LLM starts every conversation from zero. Stash gives your agent persistent memory — it remembers, recalls, consolidates, and learns across sessions. No more explaining yourself from scratch.
6
7	Open source. Self-hosted. Works with any MCP-compatible agent.
8
9	---
10
11	> Don't want to self-host?
12	> [usestash.io](https://usestash.io) is the hosted cloud version — sign in with Google, copy one MCP URL, and you're done. Free to start.
13
14	---
15
16	## Quick Start
17
18	```bash
19	git clone https://github.com/alash3al/stash.git
20	cd stash
21	cp .env.example .env # edit with your API key + model
22	docker compose up
23	```
24
25	That's it. Postgres + pgvector, migrations, MCP server with background consolidation — all in one command.
26
27	Next: [Getting Started guide](docs/GETTING_STARTED.md) — connect your MCP client, run `init` / `remember` / `recall`, and verify everything works.
28
29	Fully local (no cloud API): [Ollama setup guide](docs/LOCAL_OLLAMA.md) — host Ollama + Docker Compose, private embeddings and reasoner.
30
31	## Atlas Cloud
32
33	> [Atlas Cloud](https://www.atlascloud.ai/?utm_source=github&utm_medium=link&utm_campaign=stash) is a full-modal AI inference platform that gives developers a single AI API to access video generation, image generation, and LLM APIs. Instead of managing multiple vendor integrations, you connect once and get unified access to 300+ curated models across all modalities.
34	>
35	> Check out Atlas Cloud's coding plan promotion: [https://www.atlascloud.ai/console/coding-plan](https://www.atlascloud.ai/console/coding-plan)
36
37	Stash already supports Atlas Cloud through its OpenAI-compatible API. Set your `.env` like this:
38
39	```bash
40	STASH_OPENAI_API_KEY=your-atlas-cloud-api-key
41	STASH_OPENAI_BASE_URL=https://api.atlascloud.ai/v1
42	STASH_EMBEDDING_MODEL=text-embedding-3-small
43	STASH_REASONER_MODEL=deepseek-ai/DeepSeek-V3-0324
44	STASH_VECTOR_DIM=1536
45	```
46
47	Atlas Cloud works well here because Stash only needs:
48
49	- an embeddings model for vectorization
50	- a chat-capable reasoning model for consolidation
51	- an OpenAI-compatible base URL and API key
52
53	See [Getting Started](docs/GETTING_STARTED.md) for a fuller configuration checklist.
54
55	## MCP Client Setup
56
57	After `docker compose up`, Stash exposes an MCP server over SSE at:
58
59	```
60	http://localhost:8080/sse
61	```
62
63	Point any MCP-compatible client at that URL. Example configs:
64
65	Cursor — `~/.cursor/mcp.json`
66	```json
67	{
68	"mcpServers": {
69	"stash": {
70	"url": "http://localhost:8080/sse"
71	}
72	}
73	}
74	```
75
76	Claude Desktop — `claude_desktop_config.json`
77	```json
78	{
79	"mcpServers": {
80	"stash": {
81	"url": "http://localhost:8080/sse"
82	}
83	}
84	}
85	```
86
87	OpenCode — `~/.config/opencode/config.json`
88	```json
89	{
90	"mcp": {
91	"stash": {
92	"type": "remote",
93	"url": "http://localhost:8080/sse",
94	"enabled": true
95	}
96	}
97	}
98	```
99
100	Windsurf — `~/.codeium/windsurf/mcp_config.json`
101	```json
102	{
103	"mcpServers": {
104	"stash": {
105	"url": "http://localhost:8080/sse"
106	}
107	}
108	}
109	```
110
111	Works with any agent that supports MCP over SSE — Claude Desktop, Cursor, Windsurf, Cline, Continue, OpenAI Agents, Ollama, OpenRouter, Atlas Cloud-backed setups, and more.
112
113	## What It Does
114
115	Stash is a cognitive layer between your AI agent and the world. Episodes become facts. Facts become relationships. Relationships become patterns. Patterns become wisdom.
116
117	A 9-stage consolidation pipeline turns raw observations into structured knowledge — facts, relationships, causal links, patterns, contradictions, goal tracking, failure patterns, and hypothesis verification. Each stage only processes new data since the last run.
118
119	## Stash Cloud (Beta — Free)
120
121	A hosted, multi-tenant version of Stash is available at [usestash.io](https://usestash.io/) and is currently free while in beta.
122
123	The cloud version is written from scratch — it shares no code with this repository. It is designed from the ground up for scalability, multi-tenancy, and long-term sustainability as a product. Feature set

6 picks

Type

stash

How to install stash?

Related

stash

How to install stash?

Related