ИИ-агенты
OpenViking
A self-evolving context database unifying agent memory, RAG, and skills.
★ 39,4 тыс. звёздAGPL-3.0
Смотреть руководство →dash — проект с открытым исходным кодом из категории «ИИ-агенты» от agno-agi (agno-agi/dash на GitHub). У него 2 267 звёзд на GitHub и 253 форка, он распространяется по лицензии Apache-2.0 и последний раз обновлялся на GitHub 10 июля 2026.
Agno's self-learning data agent: grounded text-to-SQL answers over your database that improve with use.
Decide which database and business questions the agent should answer, and what knowledge (metrics definitions, gotchas, example queries) it needs.
Clone the repository and set up the Python virtual environment with the provided script.
Start the bundled database container, generate the sample data, set your OpenAI API key, and load the knowledge base.
Ask questions through the CLI mode or launch the AgentOS web UI, then run the evaluation suite to check answer quality.
Answers are grounded in multiple context layers including table schemas, business rules such as metric definitions, and known gotchas, rather than the schema alone.
A load_knowledge script upserts curated queries and rules, so the agent's accuracy improves as verified knowledge accumulates.
Five eval categories (accuracy, routing, security, governance, and boundaries) run through Agno's eval framework from the command line.
Analyst access is read-only at the PostgreSQL level, engineer writes are scoped to one schema, and destructive SQL is refused, as infrastructure guardrails rather than prompt instructions.
Run interactively in the terminal or serve the agent through AgentOS with its web UI, with role-based access control in production mode.
With a Slack bot token and signing secret, the same agent can answer data questions inside Slack.
Use the bundled venv setup script and activate the environment.
git clone https://github.com/agno-agi/dash.git
cd dash
./scripts/venv_setup.sh && source .venv/bin/activateBring up the database, generate sample data, and load the knowledge base; set OPENAI_API_KEY first.
docker compose up -d dash-db
python scripts/generate_data.py
python scripts/load_knowledge.pyUse CLI mode for local work, or AgentOS mode for the web UI.
python -m dash
python -m app.mainIt is maintained by agno-agi, the team behind the Agno agent framework.
Yes, Dash is Apache-2.0 licensed. You supply your own model access, and the documented configuration requires an OpenAI API key.
Yes for the default setup: OPENAI_API_KEY is listed as required, so usage is billed by your model provider.
Generic text-to-SQL demos generate SQL from a schema alone; Dash adds curated business knowledge, layered grounding, security enforcement at the database level, and a formal eval suite.
Clone the repo, run the venv setup, start dash-db with Docker Compose, generate data, load knowledge, and run python -m dash.
A self-evolving context database unifying agent memory, RAG, and skills.
Your self-hostable AI second brain for answers, agents, and automations.
Write coding-agent rules once, then apply them to every AI coding tool you use.
Recursive self-improving physical agent OS — verified, evidence-based execution for robots and simulators
Chat with your data: an LLM agent that builds charts and business insights.
A self-improving QA agent that writes tests in plain English.
Другие проекты из категории «ИИ-агенты» с наибольшим числом звёзд в каталоге LibreAI — реалистичный список альтернатив, если dash вам не подходит.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Соседние проекты из категории «ИИ-агенты» в каталоге, ближайшие по числу звёзд к dash — продолжайте изучать полку по обе стороны от него.
NVIDIA's object-oriented, Pythonic approach to building AI agents.
Unla MCP Gateway: turn existing APIs and MCP servers into MCP endpoints with zero code changes.
Self-managing, unbounded context and long-term memory for coding agents: one session, for life.
Self-hosted AI vulnerability research: agents find, validate, deduplicate, and rank real security issues.
A memory agent for your agent fleet: it curates, reconciles, and expires shared memories across tools.
Open Cowork: a desktop AI agent app with one-click Claude Code, MCP tools, skills, and VM sandboxing.
Бесплатно и открыто — забирайте исходный код, ставьте звезду репозиторию и начинайте создавать.