Examples - Nia AI Documentation

All examples assume you have initialized the SDK:

Python
TypeScript

import os
from nia_py.sdk import NiaSDK

sdk = NiaSDK(api_key=os.environ["NIA_API_KEY"])

import { NiaSDK } from "nia-ai-ts";

const sdk = new NiaSDK({ apiKey: process.env.NIA_API_KEY! });

Repository Operations

Index a Repository

Python
TypeScript

from nia_py import AuthenticatedClient
from nia_py.api.v2_api import index_repository_v2_v2_repositories_post
from nia_py.models import RepositoryRequest

client = AuthenticatedClient(
    base_url="https://apigcp.trynia.ai/v2",
    token=os.environ["NIA_API_KEY"],
)

result = index_repository_v2_v2_repositories_post.sync(
    client=client,
    body=RepositoryRequest(repository="vercel/ai", branch="main"),
)
print(f"Indexing started: {result}")

Or with the high-level SDK:

sdk.sources.create({"url": "https://github.com/vercel/ai", "branch": "main"})

import { V2ApiRepositoriesService } from "nia-ai-ts";

const result = await V2ApiRepositoriesService.indexRepositoryV2V2RepositoriesPost({
  repository: "vercel/ai",
  branch: "main",
});
console.log("Indexing started:", result);

Or with the high-level SDK:

await sdk.sources.create({ url: "https://github.com/vercel/ai", branch: "main" });

Check Repository Status

Python
TypeScript

from nia_py.api.v2_api import get_repository_status_v2_v2_repositories_repository_id_get

status = get_repository_status_v2_v2_repositories_repository_id_get.sync(
    client=client,
    repository_id="vercel/ai",
)
print(f"Status: {status}")

const status = await V2ApiRepositoriesService.getRepositoryStatusV2V2RepositoriesRepositoryIdGet(
  "vercel/ai"
);
console.log("Status:", status);

Browse Repository Tree

Python
TypeScript

from nia_py.api.v2_api import get_repository_tree_v2_v2_repositories_repository_id_tree_get

tree = get_repository_tree_v2_v2_repositories_repository_id_tree_get.sync(
    client=client,
    repository_id="vercel/ai",
)
print(tree)

const tree = await V2ApiRepositoriesService.getRepositoryTreeV2V2RepositoriesRepositoryIdTreeGet(
  "vercel/ai"
);
console.log(tree);

Read File Content

Python
TypeScript

from nia_py.api.v2_api import get_repository_content_v2_v2_repositories_repository_id_content_get

content = get_repository_content_v2_v2_repositories_repository_id_content_get.sync(
    client=client,
    repository_id="vercel/ai",
    path="packages/ai/src/index.ts",
)
print(content)

const content = await V2ApiRepositoriesService.getRepositoryContentV2V2RepositoriesRepositoryIdContentGet(
  "vercel/ai",
  "packages/ai/src/index.ts"
);
console.log(content);

Grep Search in Repository

Python
TypeScript

from nia_py.api.v2_api import grep_repository_v2_v2_repositories_repository_id_grep_post
from nia_py.models import CodeGrepRequest

results = grep_repository_v2_v2_repositories_repository_id_grep_post.sync(
    client=client,
    repository_id="vercel/ai",
    body=CodeGrepRequest(
        pattern="async function",
        path="src/",
        context_lines=2,
        case_sensitive=False,
    ),
)
print(results)

import { V2ApiRepositoriesService } from "nia-ai-ts";

const results = await V2ApiRepositoriesService.grepRepositoryV2V2RepositoriesRepositoryIdGrepPost(
  "vercel/ai",
  {
    pattern: "async function",
    path: "src/",
    context_lines: 2,
    case_sensitive: false,
  }
);
console.log(results);

AI Search

Semantic Code Search

Python
TypeScript

result = sdk.search.query(
    messages=[{"role": "user", "content": "How does streaming work in the AI SDK?"}],
    repositories=["vercel/ai"],
    include_sources=True,
)
print(result)

const result = await sdk.search.query({
  messages: [{ role: "user", content: "How does streaming work in the AI SDK?" }],
  repositories: ["vercel/ai"],
  include_sources: true,
});
console.log(result);

Universal Search

Searches across all your indexed repositories and documentation automatically.

Python
TypeScript

result = sdk.search.universal(
    query="How to implement tool calling?",
    top_k=10,
    include_repos=True,
    include_docs=True,
)
print(result)

const result = await sdk.search.universal({
  query: "How to implement tool calling?",
  top_k: 10,
  include_repos: true,
  include_docs: true,
});
console.log(result);

Web Search

Python
TypeScript

result = sdk.search.web(query="latest LLM developments 2025")
print(result)

const result = await sdk.search.web({ query: "latest LLM developments 2025" });
console.log(result);

Deep Research

Multi-step research with AI analysis and citations.

Python
TypeScript

result = sdk.search.deep(
    query="Compare React Server Components vs traditional SSR approaches",
    output_format="comparison table",
)
print(result)

const result = await sdk.search.deep({
  query: "Compare React Server Components vs traditional SSR approaches",
  output_format: "comparison table",
});
console.log(result);

Documentation Sources

Index Documentation

Python
TypeScript

from nia_py.api.v2_api import create_data_source_v2_v2_data_sources_post
from nia_py.models import DataSourceRequest

result = create_data_source_v2_v2_data_sources_post.sync(
    client=client,
    body=DataSourceRequest(
        url="https://docs.anthropic.com",
        display_name="Anthropic Docs",
        crawl_entire_domain=True,
    ),
)
print(f"Indexing started: {result}")

Or with the high-level SDK:

sdk.sources.create({
    "url": "https://docs.anthropic.com",
    "display_name": "Anthropic Docs",
})

import { V2ApiDataSourcesService } from "nia-ai-ts";

const result = await V2ApiDataSourcesService.createDataSourceV2V2DataSourcesPost({
  url: "https://docs.anthropic.com",
  display_name: "Anthropic Docs",
  crawl_entire_domain: true,
});
console.log("Indexing started:", result);

Or with the high-level SDK:

await sdk.sources.create({
  url: "https://docs.anthropic.com",
  display_name: "Anthropic Docs",
});

List Sources

Python
TypeScript

# List all sources
all_sources = sdk.sources.list()

# Filter by type
repos = sdk.sources.list(type="repository")
docs = sdk.sources.list(type="documentation")

# Search by name
results = sdk.sources.list(query="anthropic")

// List all sources
const allSources = await sdk.sources.list({});

// Filter by type
const repos = await sdk.sources.list({ type: "repository" });
const docs = await sdk.sources.list({ type: "documentation" });

// Search by name
const results = await sdk.sources.list({ query: "anthropic" });

Search Documentation

Python
TypeScript

result = sdk.search.query(
    messages=[{"role": "user", "content": "How to use Claude's tool use feature?"}],
    data_sources=["Anthropic Docs"],
    search_mode="sources",
)
print(result)

const result = await sdk.search.query({
  messages: [{ role: "user", content: "How to use Claude's tool use feature?" }],
  data_sources: ["Anthropic Docs"],
  search_mode: "sources",
});
console.log(result);

Research Papers

Index an arXiv Paper

Python
TypeScript

from nia_py.api.v2_api import index_research_paper_v2_v2_research_papers_post
from nia_py.models import ResearchPaperRequest

result = index_research_paper_v2_v2_research_papers_post.sync(
    client=client,
    body=ResearchPaperRequest(url="https://arxiv.org/abs/2303.08774"),
)
print(result)

import { V2ApiDataSourcesService } from "nia-ai-ts";

const result = await V2ApiDataSourcesService.indexResearchPaperV2V2ResearchPapersPost({
  url: "https://arxiv.org/abs/2303.08774",
});
console.log(result);

Package Search

Search source code of public packages on npm and PyPI without indexing.

Python
TypeScript

from nia_py.api.v2_api import package_search_hybrid_v2_v2_package_search_hybrid_post
from nia_py.models import PackageSearchHybridRequest

result = package_search_hybrid_v2_v2_package_search_hybrid_post.sync(
    client=client,
    body=PackageSearchHybridRequest(
        registry="py_pi",
        package_name="openai",
        semantic_queries=["How does streaming work?"],
    ),
)
print(result)

import { V2ApiPackageSearchService } from "nia-ai-ts";

const result = await V2ApiPackageSearchService.packageSearchHybridV2V2PackageSearchHybridPost({
  registry: "py_pi",
  package_name: "openai",
  semantic_queries: ["How does streaming work?"],
});
console.log(result);

Oracle Research

Create and Poll a Research Job

Python
TypeScript

# Start a research job
job = sdk.oracle.create_job(
    query="Compare React Server Components vs traditional SSR",
    repositories=["vercel/next.js"],
    output_format="comparison table",
    model="claude-sonnet-4-5-20250929",
)
print(f"Job ID: {job['id']}, Status: {job['status']}")

# Poll until complete (default 10min timeout)
result = sdk.oracle.wait_for_job(
    job_id=job["id"],
    timeout_seconds=300.0,
    poll_interval_seconds=3.0,
)
print(result)

// Start a research job
const job = await sdk.oracle.createJob({
  query: "Compare React Server Components vs traditional SSR",
  repositories: ["vercel/next.js"],
  output_format: "comparison table",
  model: "claude-sonnet-4-5-20250929",
});
console.log(`Job ID: ${job.id}, Status: ${job.status}`);

// Poll until complete (default 10min timeout)
const result = await sdk.oracle.waitForJob(job.id, 300_000, 3_000);
console.log(result);

Stream Research Events

Python
TypeScript

job = sdk.oracle.create_job(query="Analyze the caching strategy in Next.js App Router")

for event in sdk.oracle.stream_job_events(job_id=job["id"]):
    # Each event is a parsed JSON dict from the SSE stream
    event_type = event.get("type", "unknown")
    print(f"[{event_type}] {event}")

const job = await sdk.oracle.createJob({
  query: "Analyze the caching strategy in Next.js App Router",
});

for await (const event of sdk.oracle.streamJob(job.id)) {
  // Each event is a parsed JSON object from the SSE stream
  console.log(`[${event.type}]`, event);
}

List and Manage Jobs

Python
TypeScript

# List recent jobs
jobs = sdk.oracle.list_jobs(limit=10)

# Filter by status
completed = sdk.oracle.list_jobs(status="completed")

# Get a specific job
job = sdk.oracle.get_job(job_id="job-uuid")

import { DefaultService } from "nia-ai-ts";

// List recent jobs
const jobs = await DefaultService.listOracleJobsV2OracleJobsGet(undefined, 10);

// Get a specific job
const job = await DefaultService.getOracleJobV2OracleJobsJobIdGet("job-uuid");

Save and retrieve context across sessions.

Python
TypeScript

from nia_py.api.v2_api import (
    save_context_v2_v2_contexts_post,
    semantic_search_contexts_v2_v2_contexts_semantic_search_get,
)
from nia_py.models import ContextSaveRequest

# Save context
save_context_v2_v2_contexts_post.sync(
    client=client,
    body=ContextSaveRequest(
        title="Auth implementation notes",
        summary="Implemented OAuth2 flow with refresh tokens",
        content="Detailed implementation notes...",
        tags=["auth", "oauth"],
    ),
)

# Search contexts by semantic similarity
results = semantic_search_contexts_v2_v2_contexts_semantic_search_get.sync(
    client=client,
    query="authentication",
)
print(results)

import { V2ApiContextsService } from "nia-ai-ts";

// Save context
await V2ApiContextsService.saveContextV2V2ContextsPost({
  title: "Auth implementation notes",
  summary: "Implemented OAuth2 flow with refresh tokens",
  content: "Detailed implementation notes...",
  tags: ["auth", "oauth"],
});

// Search contexts by semantic similarity
const results = await V2ApiContextsService.semanticSearchContextsV2V2ContextsSemanticSearchGet(
  "authentication"
);
console.log(results);

GitHub Tracer

Live code search across GitHub repositories without indexing.

Python
TypeScript

from nia_py.api.v2_api import create_tracer_job_v2_github_tracer_post
from nia_py.models import TracerRequest

job = create_tracer_job_v2_github_tracer_post.sync(
    client=client,
    body=TracerRequest(
        query="How does the streaming API work?",
        repositories=["vercel/ai"],
    ),
)
print(f"Tracer job: {job}")

import { GithubSearchService } from "nia-ai-ts";

const job = await GithubSearchService.createTracerJobV2GithubTracerPost({
  query: "How does the streaming API work?",
  repositories: ["vercel/ai"],
});
console.log("Tracer job:", job);

Async Support (Python)

Every low-level API function has four variants:

from nia_py.api.v2_api import search_universal_v2_v2_search_universal_post
from nia_py.models import UniversalSearchRequest
from nia_py.types import Response

body = UniversalSearchRequest(query="test")

# Sync — returns parsed data or None
result = search_universal_v2_v2_search_universal_post.sync(client=client, body=body)

# Sync detailed — returns Response with status_code, headers, parsed
response: Response = search_universal_v2_v2_search_universal_post.sync_detailed(client=client, body=body)

# Async — returns parsed data or None
result = await search_universal_v2_v2_search_universal_post.asyncio(client=client, body=body)

# Async detailed — returns Response
response = await search_universal_v2_v2_search_universal_post.asyncio_detailed(client=client, body=body)

SDKs

​Repository Operations

​Index a Repository

​Check Repository Status

​Browse Repository Tree

​Read File Content

​Grep Search in Repository

​AI Search

​Semantic Code Search

​Universal Search

​Web Search

​Deep Research

​Documentation Sources

​Index Documentation

​List Sources

​Search Documentation

​Research Papers

​Index an arXiv Paper

​Package Search

​Oracle Research

​Create and Poll a Research Job

​Stream Research Events

​List and Manage Jobs

​Context Sharing

​GitHub Tracer

​Async Support (Python)

Repository Operations

Index a Repository

Check Repository Status

Browse Repository Tree

Read File Content

Grep Search in Repository

AI Search

Semantic Code Search

Universal Search

Web Search

Deep Research

Documentation Sources

Index Documentation

List Sources

Search Documentation

Research Papers

Index an arXiv Paper

Package Search

Oracle Research

Create and Poll a Research Job

Stream Research Events

List and Manage Jobs

Context Sharing

GitHub Tracer

Async Support (Python)