raysurfer 1.4.0

AI maintained skills for vertical agents

RaySurfer Python SDK

Website · Docs · Dashboard

AI Maintained Skills for Vertical Agents. Re-use verified code from prior runs rather than serial tool calls or generating code per execution.

Installation

pip install raysurfer

Setup

Set your API key:

export RAYSURFER_API_KEY=your_api_key_here

Get your key from the dashboard.

Low-Level API

For custom integrations, use the RaySurfer client directly with any LLM provider.

Complete Example with Anthropic API

import anthropic
from raysurfer import RaySurfer
from raysurfer.types import FileWritten, LogFile

client = RaySurfer(api_key="your_raysurfer_api_key")
task = "Fetch GitHub trending repos"

# 1. Search for cached code matching a task
result = client.search(
    task=task,
    top_k=5,
    min_verdict_score=0.3,
)

for match in result.matches:
    print(f"{match.code_block.name}: {match.combined_score}")
    print(f"  Source: {match.code_block.source[:80]}...")

# 2. Upload a new code file after execution
file = FileWritten(path="fetch_repos.py", content="def fetch(): ...")
client.upload(
    task=task,
    file_written=file,
    succeeded=True,
    execution_logs="Fetched 10 trending repos successfully",
    dependencies={"httpx": "0.27.0", "pydantic": "2.5.0"},
)

# 2b. Bulk upload prompts/logs/code for sandboxed grading
logs = [LogFile(path="logs/run.log", content="Task completed", encoding="utf-8")]
client.upload_bulk_code_snips(
    prompts=["Build a CLI tool", "Add CSV support"],
    files_written=[FileWritten(path="cli.py", content="def main(): ...")],
    log_files=logs,
)

# 3. Vote on whether a cached snippet was useful
client.vote_code_snip(
    task=task,
    code_block_id=result.matches[0].code_block.id,
    code_block_name=result.matches[0].code_block.name,
    code_block_description=result.matches[0].code_block.description,
    succeeded=True,
)

Async Version

import anthropic
from raysurfer import AsyncRaySurfer
from raysurfer.types import FileWritten

async with AsyncRaySurfer(api_key="your_api_key") as client:
    # 1. Search for cached code
    result = await client.search(task="Fetch GitHub trending repos")

    for match in result.matches:
        print(f"{match.code_block.name}: {match.combined_score}")

    # 2. Upload a new code file after execution
    file = FileWritten(path="fetch_repos.py", content="def fetch(): ...")
    await client.upload(
        task="Fetch GitHub trending repos",
        file_written=file,
        succeeded=True,
        execution_logs="Fetched 10 trending repos successfully",
    )

    # 3. Vote on snippet manually
    await client.vote_code_snip(
        task="Fetch GitHub trending repos",
        code_block_id=result.matches[0].code_block.id,
        code_block_name=result.matches[0].code_block.name,
        code_block_description=result.matches[0].code_block.description,
        succeeded=True,
    )

Client Options

client = RaySurfer(
    api_key="your_api_key",
    base_url="https://api.raysurfer.com",  # optional
    timeout=30,                             # optional, in seconds
    organization_id="org_xxx",              # optional, for team namespacing
    workspace_id="ws_xxx",                  # optional, for enterprise namespacing
    snips_desired="company",                # optional, snippet scope
    public_snips=True,                      # optional, include community snippets
)

Response Fields

The search() response includes:

Field	Type	Description
matches	list[SearchMatch]	Matching code blocks with scoring
total_found	int	Total matches found
cache_hit	bool	Whether results were from cache

Each SearchMatch contains code_block (with id, name, source, description, entrypoint, language, dependencies), combined_score, vector_score, verdict_score, thumbs_up, thumbs_down, filename, and entrypoint.

Store a Code Block with Full Metadata

result = client.store_code_block(
    name="GitHub User Fetcher",
    source="def fetch_user(username): ...",
    entrypoint="fetch_user",
    language="python",
    description="Fetches user data from GitHub API",
    tags=["github", "api", "user"],
    dependencies={"httpx": "0.27.0", "pydantic": "2.5.0"},
)

Retrieve Few-Shot Examples

examples = client.get_few_shot_examples(task="Parse CSV files", k=3)

for ex in examples:
    print(f"Task: {ex.task}")
    print(f"Code: {ex.code_snippet}")

Retrieve Task Patterns

patterns = client.get_task_patterns(
    task="API integration",
    min_thumbs_up=5,
    top_k=20,
)

for p in patterns:
    print(f"{p.task_pattern} -> {p.code_block_name}")

User-Provided Votes

Instead of relying on AI voting, provide your own votes:

# Single upload with your own vote (AI voting is skipped)
client.upload(
    task="Fetch GitHub trending repos",
    file_written=file,
    succeeded=True,
    user_vote=1,  # 1 = thumbs up, -1 = thumbs down
)

# Bulk upload with per-file votes (AI grading is skipped)
client.upload_bulk_code_snips(
    prompts=["Build a CLI tool", "Add CSV support"],
    files_written=files,
    log_files=logs,
    user_votes={
        "app.py": 1,     # thumbs up
        "utils.py": -1,  # thumbs down
    },
)

Method Reference

Method	Description
search(task, top_k, min_verdict_score, prefer_complete, input_schema)	Search for cached code snippets
get_code_snips(task, top_k, min_verdict_score)	Retrieve cached code snippets by semantic search
retrieve_best(task, top_k, min_verdict_score)	Retrieve the single best match
get_few_shot_examples(task, k)	Retrieve few-shot examples for code generation prompting
get_task_patterns(task, min_thumbs_up, top_k)	Retrieve proven task-to-code mappings
store_code_block(name, source, entrypoint, language, description, tags, dependencies, ...)	Store a code block with full metadata
upload(task, file_written, succeeded, use_raysurfer_ai_voting, user_vote, execution_logs, dependencies)	Store a single code file with optional dependency versions
upload_bulk_code_snips(prompts, files_written, log_files, use_raysurfer_ai_voting, user_votes)	Bulk upload for grading (AI votes by default, or provide per-file votes)
delete(snippet_id)	Delete a snippet by ID or name
vote_code_snip(task, code_block_id, name, description, succeeded)	Vote on snippet usefulness

Exceptions

Both sync and async clients include built-in retry logic with exponential backoff for transient failures (429, 5xx, network errors).

Exception	Description
RaySurferError	Base exception for all Raysurfer errors
APIError	API returned an error response (includes status_code)
AuthenticationError	API key is invalid or missing
CacheUnavailableError	Cache backend is unreachable
RateLimitError	Rate limit exceeded after retries (includes retry_after)
ValidationError	Request validation failed (includes field)

from raysurfer import RaySurfer
from raysurfer.exceptions import RateLimitError

client = RaySurfer(api_key="your_api_key")

try:
    result = client.get_code_snips(task="Fetch GitHub repos")
except RateLimitError as e:
    print(f"Rate limited after retries: {e}")
    if e.retry_after:
        print(f"Try again in {e.retry_after}s")

---

Claude Agent SDK Drop-in

Swap your client class and method names. Options come directly from claude_agent_sdk:

# Before
from claude_agent_sdk import ClaudeSDKClient, ClaudeAgentOptions

# After
from raysurfer import RaysurferClient
from claude_agent_sdk import ClaudeAgentOptions

options = ClaudeAgentOptions(
    allowed_tools=["Read", "Write", "Bash"],
    system_prompt="You are a helpful assistant.",
)

async with RaysurferClient(options) as client:
    await client.query("Generate quarterly report")
    async for msg in client.response():
        print(msg)

Method Mapping

Claude SDK	Raysurfer
ClaudeSDKClient(options)	RaysurferClient(options)
await client.query(prompt)	await client.query(prompt)
client.receive_response()	client.response()

Full Example

import asyncio
import os
from raysurfer import RaysurferClient
from claude_agent_sdk import ClaudeAgentOptions

os.environ["RAYSURFER_API_KEY"] = "your_api_key"

async def main():
    options = ClaudeAgentOptions(
        allowed_tools=["Read", "Write", "Bash"],
        system_prompt="You are a helpful assistant.",
    )

    async with RaysurferClient(options) as client:
        # First run: generates and caches code
        await client.query("Fetch GitHub trending repos")
        async for msg in client.response():
            print(msg)

        # Second run: retrieves from cache (instant)
        await client.query("Fetch GitHub trending repos")
        async for msg in client.response():
            print(msg)

asyncio.run(main())

Without Caching

If RAYSURFER_API_KEY is not set, RaysurferClient behaves exactly like ClaudeSDKClient — no caching, just a pass-through wrapper.

Snippet Retrieval Scope

Control which cached snippets are retrieved using snips_desired:

from raysurfer import RaysurferClient
from claude_agent_sdk import ClaudeAgentOptions

options = ClaudeAgentOptions(
    allowed_tools=["Read", "Write", "Bash"],
)

# Include company-level snippets
client = RaysurferClient(
    options,
    snips_desired="company",  # Company-level snippets (Team/Enterprise)
)

# Enterprise: Retrieve client-specific snippets only
client = RaysurferClient(
    options,
    snips_desired="client",   # Client workspace snippets (Enterprise only)
)

Configuration	Required Tier
snips_desired="company"	TEAM or ENTERPRISE
snips_desired="client"	ENTERPRISE only

Public Snippets

Include community public snippets (crawled from GitHub) in retrieval results alongside your private snippets:

# High-level
client = RaysurferClient(options, public_snips=True)

# Low-level
client = RaySurfer(api_key="...", public_snips=True)

Programmatic Tool Calling

Register local tools, then either:

1. pass in user_code (primary mode), or
2. generate code inside the sandbox with your own provider key + prompt (optional mode).

import asyncio
from raysurfer import AsyncRaySurfer

async def main():
    rs = AsyncRaySurfer(api_key="your_api_key")

    @rs.tool
    def add(a: int, b: int) -> int:
        """Add two numbers together."""
        return a + b

    @rs.tool
    def multiply(a: int, b: int) -> int:
        """Multiply two numbers together."""
        return a * b

    user_code = """
intermediate = add(5, 3)
final = multiply(intermediate, 2)
print(final)
"""
    result = await rs.execute(
        "Add 5 and 3, then multiply the result by 2",
        user_code=user_code,
    )
    print(result.result)       # "16"
    print(result.tool_calls)   # [ToolCallRecord(tool_name='add', ...), ToolCallRecord(tool_name='multiply', ...)]
    print(result.cache_hit)    # False (reserved field for execute)

asyncio.run(main())

The @rs.tool decorator introspects your function signature to build a JSON schema. Both sync and async callbacks are supported.

How It Works

1. SDK connects a WebSocket to the server for tool call routing
2. Your app sends either user_code (primary mode) or codegen_* inputs (optional mode) to /api/execute/run
3. Code runs in a sandboxed environment — tool calls are routed back to your local functions via WebSocket
4. Results are returned with full tool call history

Execute Options

result = await rs.execute(
    "Your task description",
    user_code="print(add(1, 2))",  # Primary mode
    timeout=300,                    # Max execution time in seconds (default 300)
)

# Optional mode: generate code in sandbox using your own key + prompt
result = await rs.execute(
    "Your task description",
    codegen_api_key="your_anthropic_key",
    codegen_prompt="Write Python code that uses add(a, b) and prints the result for 2 + 3.",
    codegen_model="claude-opus-4-6",
)

ExecuteResult Fields

Field	Type	Description
execution_id	str	Unique execution identifier
result	str | None	Stdout output from the script
exit_code	int	Process exit code (0 = success)
duration_ms	int	Total execution time
cache_hit	bool	Reserved field (currently always False for execute)
error	str | None	Error message if exit_code != 0
tool_calls	list[ToolCallRecord]	All tool calls made during execution

Agent-Accessible Functions

Mark functions as agent-callable with @agent_accessible(). Metadata (name, description, input schema) is auto-inferred from the function signature and docstring:

from raysurfer import agent_accessible, to_anthropic_tool

@agent_accessible()
def fetch_user(username: str, include_repos: bool = False):
    """Fetch a GitHub user profile.

    Args:
        username: The GitHub username to look up.
        include_repos: Whether to include the user's repositories.
    """
    ...

# Override any auto-inferred field
@agent_accessible(description="Custom description", name="custom_name")
def another_function(x: int):
    ...

Convert to Anthropic Tool Format

tool_def = to_anthropic_tool(fetch_user)
# {"name": "fetch_user", "description": "Fetch a GitHub user profile.", "input_schema": {...}}

Org/Workspace Scoping

Scope functions to a specific organization or workspace for upload:

@agent_accessible(org_id="org_xxx", workspace_id="ws_xxx")
def scoped_function(data: str):
    """Only visible within the specified workspace."""
    ...

License

MIT