Practical overview
When it fits
Use SAM when developing a Python agent that needs more than a model call: persistent context, tool execution, event handling, provider choice, or connections to Solana-related workflows.
Who it is for
- Python developers building tool-using agents for Solana or DeFi workflows.
- Teams that need agent sessions, event streaming, provider switching or custom tools in a common runtime.
- Developers exploring wallet-aware agents who want policy-driven tool access and validation mechanisms.
Documented uses
- Build an agent that calls tools and retains context across persistent SQLite-backed sessions.
- Connect an agent to Solana wallet, balance, market-data or transaction-validation workflows.
- Switch among supported language-model providers and compatible local endpoints.
- Expose agent responses through the CLI, HTTP API or Streamlit example interface.
- Extend the tool set through plugins without replacing the core agent loop.
Requirements & limits
The project is provided as a publicly accessible GitHub repository. Its README states an MIT license, while the repository metadata supplied here does not list a license.
- SAM is a framework for developers to build and configure agents, rather than a hosted end-user agent service.
- The README names OpenAI, Anthropic, xAI and local-compatible endpoints as model-provider options.
- Persistent sessions use SQLite; wallet and key storage is described as encrypted with Fernet and the OS keyring.
- Wallet, market and transaction workflows remain operationally consequential: the project says its controls do not make automated trading safe by magic.
- The repository's documented setup uses Git and uv and calls for provider credentials and a Fernet key in manual configuration.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Install SAM and complete guided onboarding
Set up the repository and run its guided configuration flow.
Before you start
- Git
- uv
- Provider credentials and a Fernet key if configuring manually
Steps
Clone the repository with the command below.
git clone https://github.com/prfagit/sam-frameworkChange into the
sam-frameworkdirectory.Install the project dependencies with the command below.
uv syncRun the command below to start guided setup.
uv run sam onboard
Expected outcome
The repository is installed and the guided onboarding flow is started to write local configuration.
Guide limitations
- The documented steps cover installation and onboarding, not a specific agent workflow.
- Manual configuration lists LLM provider keys and a SAM Fernet key; Brave Search is also listed as an optional configuration key.
Start an interactive agent
Launch SAM's documented interactive agent after setup.
Before you start
- SAM installed with
uv sync - Local configuration completed through onboarding or manual setup
Steps
From the repository directory, run the command below.
uv run sam runTo use a named session, run the command below.
uv run sam run --session trading_session
Expected outcome
The CLI starts an interactive agent, with the named-session form available for a session called `trading_session`.
Guide limitations
- The commands document launching an agent, not a particular prompt, tool action or trading result.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 17 October 2026
- Original website title
- prfagit/sam-framework · Official homepage title · retained in its source language