YantrikDB Server

https://github.com/yantrikos/yantrikdb-server
Admin
YantrikDB Server is a multi-tenant cognitive memory database server that provides memory decay,...
Tokens:182,272
Snippets:1,345
Trust Score:6.8
License:Apache-2.0
Update:1 month ago
Show doc for...
Tokens:
Raw
### Quick start server setup

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/crates/yantrikdb-server/README.md

Initialize a database, generate an access token, and start the server.

```bash
yantrikdb db --data-dir ./data create default
yantrikdb token --data-dir ./data create --db default
yantrikdb serve --data-dir ./data
# Wire protocol on :7437, HTTP on :7438
```

--------------------------------

### Download and Setup YantrikDB Server

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/SERVER_README.md

Commands to download the YantrikDB binaries, set up a database directory, generate an access token, and start the server.

```bash
# Download
wget https://github.com/yantrikos/yantrikdb-server/releases/latest/download/yantrikdb-linux-amd64
chmod +x yantrikdb-linux-amd64
mv yantrikdb-linux-amd64 /usr/local/bin/yantrikdb

wget https://github.com/yantrikos/yantrikdb-server/releases/latest/download/yql-linux-amd64
chmod +x yql-linux-amd64
mv yql-linux-amd64 /usr/local/bin/yql

# Setup
yantrikdb db --data-dir ./data create default
TOKEN=$(yantrikdb token --data-dir ./data create --db default | grep ^ydb_)
echo "$TOKEN" > token.txt

# Run
yantrikdb serve --data-dir ./data
# Wire protocol on :7437, HTTP on :7438
```

--------------------------------

### Install YantrikDB library

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/README.md

Installation commands for Python and Rust environments.

```bash
pip install yantrikdb
# or
cargo add yantrikdb
```

--------------------------------

### Install yantrikdb-witness via Cargo

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/crates/yantrikdb-witness/README.md

Use the cargo package manager to install the witness daemon.

```bash
cargo install yantrikdb-witness
```

--------------------------------

### Install YQL via Cargo

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/crates/yql/README.md

Install the YQL CLI tool using the Rust package manager.

```bash
cargo install yql
```

--------------------------------

### Install yantrikdb-server via Cargo

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/crates/yantrikdb-server/README.md

Install the server directly using the Rust package manager.

```bash
cargo install yantrikdb-server
```

--------------------------------

### YQL Interactive Session Example

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/crates/yql/README.md

An example session demonstrating memory storage, semantic recall, graph relationship creation, and meta-commands.

```text
yql connected to http://localhost:7438
type \h for help, \q to exit

yantrikdb> remember "Alice leads engineering at Acme" importance=0.9 domain=work
✓ stored: 019d623a-3d70-712e-9315-e1da5ee41114

yantrikdb> recall who leads engineering top=5
+---+-------+---------------------------------+--------+--------------------------------+
| # | score | text                            | domain | why                            |
+---+-------+---------------------------------+--------+--------------------------------+
| 1 | 1.41  | Alice leads engineering at Acme | work   | semantically similar (0.54)... |
+---+-------+---------------------------------+--------+--------------------------------+
(1 rows)

yantrikdb> relate Alice -> Acme as works_at
✓ edge: 019d623a-41cf-71a2 (Alice -[works_at]-> Acme)

yantrikdb> \stats
yantrikdb> \cluster
yantrikdb> \q
```

--------------------------------

### GET /v1/skills/{skill_id} response

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/docs/rfcs/rfc_022_skill_substrate_and_ryw.md

The 200 OK response containing skill details.

```json
{
  "rid":          "019de580-1234-...",
  "skill_id":     "skill.invoice.validation.v3",
  "body":         "When validating an invoice, ...",
  "applies_to":   ["invoice", "approval"],
  "skill_type":   "rule",
  "metadata":     { ... },
  "namespace":    "skill_substrate",
  "memory_type":  "procedural",
  "created_at":   1762012345.123
}
```

--------------------------------

### HTTP /v1/lookup request and response

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/docs/rfcs/rfc_022_skill_substrate_and_ryw.md

Example of a POST request to the /v1/lookup endpoint and the corresponding 200 OK JSON response.

```http
POST /v1/lookup
Authorization: Bearer <token>
Content-Type: application/json

{
  "namespace": "skill_substrate",
  "field":     "skill_id",
  "value":     "skill.invoice.validation.v3"
}
```

```json
{
  "found":  true,
  "memory": {
    "rid":         "019de580-...",
    "text":        "...",
    "metadata":    { ... },
    "namespace":   "skill_substrate",
    "memory_type": "procedural",
    "created_at":  1762012345.123
  },
  "index_used": "skill_id"
}
```

--------------------------------

### GET /v1/skills/{skill_id} request

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/docs/rfcs/rfc_022_skill_substrate_and_ryw.md

Retrieves a specific skill by its ID.

```http
GET /v1/skills/{skill_id}
Authorization: Bearer <token>
```

--------------------------------

### POST /v1/recall Request and Response

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/docs/rfcs/rfc_022_skill_substrate_and_ryw.md

Example of a recall request with a where-clause prefilter and the corresponding explain response from the server.

```json
POST /v1/recall
Authorization: Bearer <token>
Content-Type: application/json

{
  "namespace": "skill_substrate",
  "query":     "How should I validate an invoice?",
  "top_k":     5,
  "where": {
    "all": [
      { "field": "record_type", "op": "eq",       "value": "skill" },
      { "field": "skill_type",  "op": "eq",       "value": "rule" },
      { "field": "applies_to",  "op": "contains", "value": "invoice" }
    ]
  },
  "explain": true
}
```

```json
{
  "results": [ ... ],
  "plan": {
    "filter_mode":              "indexed_prefilter",
    "indexes_used":             ["record_type", "skill_type", "applies_to"],
    "candidate_count":          384,
    "vector_strategy":          "exact_rerank_over_candidates",
    "estimated_distance_computations": 384,
    "fallback_used":            false
  }
}
```

--------------------------------

### Extended /v1/health response structure

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/docs/rfcs/rfc_010_pr6_write_path_migration.md

Example JSON response for the health endpoint, including new fields for replication lag, commit log status, and last applied timestamps.

```json
{
  "status": "healthy",
  "raft_mode": "openraft",
  "cluster": {
    "node_id": 2,
    "leader_id": 4,
    "term": 5,
    "last_log_index": 1842,
    "last_applied": 1842,
    "members": [...],
    "replication_lag_log_entries": 0,        // NEW
    "last_applied_at_unix_micros": 1779983412345678,  // NEW
    "commit_log_max_high_watermark": 39427    // NEW
  },
  "single_node": null
}
```

--------------------------------

### Download pre-built binary

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/crates/yantrikdb-server/README.md

Download and prepare the latest pre-built binary for Linux.

```bash
wget https://github.com/yantrikos/yantrikdb-server/releases/latest/download/yantrikdb-linux-amd64
chmod +x yantrikdb-linux-amd64
```

--------------------------------

### Reproduce skill recall benchmark

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/benchmarks/skill_recall/RESULTS.md

Shell commands to set up the environment, generate data, seed the database, and run the benchmark scripts.

```bash
cd benchmarks/skill_recall
docker compose up -d

# Mint a token (the running server auto-creates a 'default' database).
docker exec yantrikdb-skill-bench yantrikdb token \
  --data-dir /var/lib/yantrikdb create --db default --label bench
# (capture printed token)
export YDB_TOKEN=ydb_...

python generate.py 5000
python seed.py    --nodes 127.0.0.1 --port 17438 --rate-per-sec 50
python bench.py   --nodes 127.0.0.1 --port 17438
python cache_bench.py --nodes 127.0.0.1 --port 17438

docker compose down -v   # cleanup
```

--------------------------------

### Conflict Detection Example

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/README.md

Example of how YantrikDB represents a contradiction between two memory records.

```text
"works at Google" (recorded Jan 15) vs. "works at Meta" (recorded Mar 1)
→ Conflict: identity_fact, priority: high, strategy: ask_user
```

--------------------------------

### CLI Commands

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/DESIGN.md

Common CLI commands for starting the server, managing databases, and handling data exports/imports.

```bash
yantrikdb serve                          # Start server
yantrikdb serve --config ./yantrikdb.toml
yantrikdb db create <name>               # Create database
yantrikdb db list                        # List databases
yantrikdb token create --db <name>       # Generate token
yantrikdb export <db> > backup.jsonl     # Export
yantrikdb import <db> < backup.jsonl     # Import
```

--------------------------------

### Create Database on Voter Nodes

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/SERVER_README.md

Initialize the database on each voter node. The witness node does not require this step.

```bash
yantrikdb db --data-dir ./data create default
```

--------------------------------

### Initialize cluster node

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/crates/yantrikdb-server/README.md

Initialize a cluster node configuration and start the server using the generated TOML file.

```bash
yantrikdb cluster init \
  --node-id 1 \
  --output ./node1.toml \
  --peers 192.168.1.2:7440 \
  --witnesses 192.168.1.3:7440
yantrikdb serve --config ./node1.toml
```

--------------------------------

### Start Voter Nodes

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/SERVER_README.md

Start the YantrikDB service on each voter node using its respective configuration file.

```bash
yantrikdb serve --config ./node1.toml
```

```bash
yantrikdb serve --config ./node2.toml
```

--------------------------------

### Reproduce Benchmarks

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/docs/papers/zenodo-bundle/README.md

Commands to clone the YantrikDB server repository, install necessary dependencies, and execute the benchmark scripts.

```bash
# Clone the YantrikDB server repository
git clone https://github.com/yantrikos/yantrikdb-server.git
cd yantrikdb-server/benchmarks/skill_recall

# Install dependencies
pip install tiktoken matplotlib

# Re-run all benchmarks
python token_cost_bench.py
python kill_test_ablation.py
python k_sensitivity_and_tokenizer_bench.py
python hallucination_admission_bench.py
```

--------------------------------

### Start yql interactive client

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/SERVER_README.md

Connect to the YantrikDB server using the yql command-line interface.

```bash
yql --host 192.168.4.180 -p 7438 -t ydb_cluster_3a8f...
```

--------------------------------

### Install Python SDK dependencies

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/SERVER_README.md

Install the required httpx and msgpack packages to use the YantrikDB Python SDK.

```bash
pip install httpx msgpack
# Use sdk/python/yantrikdb (or vendor it into your project)
```

--------------------------------

### Start SSE transport server

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/README.md

Command to start the MCP server using the SSE transport protocol on a specified port.

```bash
yantrikdb-mcp --transport sse --port 8420
```

--------------------------------

### Install YantrikDB MCP package

Source: https://github.com/yantrikos/yantrikdb-server/blob/main/README.md

Install the required Python package to enable YantrikDB memory tools in your MCP-compatible AI agent.

```bash
pip install yantrikdb-mcp
```
Tokens:
Raw