awresearch

Docs · Source · pip install awresearch · The Aither World

The Aither World is an operating system for agents — a Linux you can hand to one, the runtimes it works in, and the tools it works with. awnix is the Linux underneath it; awresearch is one of its 33 bricks — each installs on its own, runs offline, and needs no account.

Start here: Ask one question and get back a report where every claim carries the source it came from.

Ask a research question, get a cited report you can check.

A research agent that actually reads its sources — it searches, retrieves the real pages, extracts specific claims with citations, and refuses to state anything it didn't retrieve. Every claim in the report is traced back to where it came from.

This solves the plausible-document problem: an LLM asked to research something returns fluent, confident text whose claims were never verified against real sources. awresearch trades speed for honesty — it costs real API calls and real retrieval time, and you get a report where checking the facts is possible.

Install

pip install awresearch

Requires Python 3.10+, awdk (the agent engine), and an LLM provider (Anthropic / OpenAI / DeepSeek / local Ollama).

Quick start

from awdk.agent import AitherAgent
from awdk.identity import Identity
from awdk.memory import Memory
from awresearch.api import Researcher, Report

# Set up your LLM key in the environment (ANTHROPIC_API_KEY, OPENAI_API_KEY, etc.)

# Create a Researcher with an awdk agent
identity = Identity(name="researcher", description="Research analyst")
memory = Memory(db_path=".data/researcher.db", agent_name="researcher")
agent = AitherAgent(name="researcher", identity=identity, memory=memory)

researcher = Researcher(llm_backend=agent.llm)

# Ask a question
report = await researcher.research(
    "What are the latest breakthroughs in solid-state batteries?",
    depth="standard",
)

# Check the results
print(f"Claims: {len(report.claims)}")
print(f"Sources: {len(report.sources)}")
print(f"All claims sourced? {all(c.is_sourced for c in report.claims)}")

# Export as Markdown with citations
markdown = report.markdown()
print(markdown)

# Or as structured data
import json
print(json.dumps(report.to_dict(), indent=2))

API

Researcher(llm_backend, search_backend=None, artifacts_dir=None)

Configurable research agent.

Args:

• llm_backend — an awdk LLMRouter or compatible (required)
• search_backend — callable returning search results. Defaults to AitherSearch (multi-engine web search)
• artifacts_dir — where to store temporary files. Defaults to AITHER_DATA_DIR env var

async Researcher.research(question, depth="standard", max_sources=10) -> Report

Research a question and return a cited report.

Args:

• question — the research question
• depth — "standard" (one pass) or "deep" (iterative multi-angle)
• max_sources — maximum sources to fetch (default: 10)

Report

Result of a research session: claims, sources, validation.

Attributes:

• question — the original question
• claims — list of Claim objects
• sources — list of Source objects (each referenced by claims)
• research_depth — "standard" or "deep"

Methods:

• validate() → list of validation issues (empty = valid)
• to_dict() → JSON-serializable dict
• markdown() → Markdown report with citations

Claim

A single claim in the report.

Attributes:

• text — the claim
• sources — list of 1-based indices into Report.sources
• is_sourced — True if has at least one source
• unsourced_reason — if unsourced, why (e.g., "too general to cite")

Source

A retrieved source: URL, title, freshness/trust metadata.

Attributes:

• url — the source URL
• title — page title
• retrieved_at — ISO timestamp
• domain — domain name
• authority — domain authority score (0-1)
• freshness — freshness score (0-1, based on publication date)
• trust — combined trust score

Configuration

Set via environment variables:

• AWRESEARCH_SEARCH_BACKEND — search engine ("ddgs", "searxng", etc.). Default: multi-engine (AitherSearch)
• AITHER_DATA_DIR — where to store artifacts. Default: temp directory

Limitations

• Real API calls. Every search, every page fetch costs an HTTP request. Research is not free.
• Sources can be wrong. This tool cross-checks claims against retrieved pages, but cannot verify that pages themselves are accurate. It is not a fact-checker.
• Paywalls and access. Some sources are behind paywalls or require authentication. The tool retrieves what is publicly accessible.
• Speed. Real research takes time. A deep-research turn may take 30–60 seconds per question.

CLI

# Not yet implemented. See awresearch Python API above.
awresearch --question "Research question here" --output markdown

Architecture

awresearch is built on:

• awdk — the ReAct agent loop, memory system, and LLM routing
• AitherSearch — multi-engine web search (DuckDuckGo, SearXNG, Startpage)
• Scholarly sources — arXiv, Semantic Scholar, Crossref (for academic papers)
• Trust scoring — domain authority and content freshness detection
• Knowledge graph — awdk's GraphMemory for conversation memory and fact recall

License

Apache-2.0. See LICENSE.

Contributing

Contributions welcome. See GitHub.

---

Built by Aitherium. GitHub · Docs

The aw family

Standalone tools that share one idea: replace something you would otherwise have to trust with something you can check.

Each installs on its own, works offline, and needs no account.

	instead of trusting	you check
awdk	a framework's idea of how your agents should run	one loop you can read, pointed at a backend you already pay for
awskills	that an agent knows your procedure	the procedure written down, versioned, and loadable by any agent
awm	that memory stayed in its lane	tenant:user:project scopes, so a write cannot cross a boundary
awnode	a vendor's cloud with every prompt	a local gateway routing to backends you chose
awgraph	that grep found everything	an AST + tree-sitter call graph an agent can traverse
awgit	that no one else is editing this file	a lease, refused at commit time if you do not hold it
awseal	that the artifact came from who you think	an Ed25519 seal — the key that verifies is not the key that forges
awshare	that the download is intact	content-addressed bundles, verified on fetch
awnest	that there is a person on the other end	a verdict with evidence, where "we could not tell" is not "yes"
awnboard	a share link anyone who sees it can use	an invitation addressed to one person, for one gate, revocable
awnix	that the box is what you left it as	an immutable image you built, with atomic rollback
awrecover	that the restore worked	a restore that fully lands or does not land at all
awkno	that the docs site is up, or that you remember the family	the whole ecosystem in your terminal, with no network at all
awrelay	a SaaS in the middle of your agents	findings, alerts and coordination over your own transport
awmail	a mailbox somebody else can read	mail your agents send and receive over your own server
awfind	one vendor's idea of the web	results from whichever providers you configured
awbrowse	that the page said what you were told	the render, the DOM and the requests it made
aitherkvcache	a vendor's quantisation defaults	sub-byte KV cache kernels you can benchmark yourself
AitherZero	a pile of scripts nobody has numbered	numbered, discoverable automation with declarative playbooks
AitherConnect	what a page tells your browser to do	a federated search and desktop bridge you host
awreason	a confident paragraph	the phases it went through, and every tool call it made to get there
awrecurse	that everything you pasted in was actually read	which slices it opened, and what it concluded from each
awprism	the first explanation that fits	the ranked alternatives, and the observation that separates them
awrepl	what the agent believes the value is	the value, printed from the live session
awresearch (you are here)	a summary of pages nobody opened	every claim against the source it came from
awpredict	a model because it trained without erroring	its prediction against a self-updating lookup, on the rows that are actually novel
awkno	that the docs site is up, or that you remember the family	the whole ecosystem in your terminal, with no network at all

awnix is the ground floor — A Linux you can hand to an agent — immutable base, capabilities included.

The Aitherium ecosystem

Every repository here is public. Each publishes an aither-manifest.json beside its page, so any surface can read every sibling's — the network is browsable from any node in it.

repo	what it is	pages
awdk	Build AI agent fleets — 3 lines, any backend, local or cloud	docs
awskills	Portable agent skills — self-contained procedures an agent loads on demand	docs
awm	A portable, scoped agent memory	docs
awnode	A lightweight local gateway — bridges your apps to the AI backends you chose	docs
awrun	A priority-aware queue and dispatcher for agentic runs and ad-hoc CI builds	docs
awgraph	A semantic code graph for agents — AST + tree-sitter, call graphs	docs
awgit	Semantic version control on top of git — edit-ops and leases	docs
awseal	Sign an artifact so a stranger can verify it	docs
awshare	Publish an artifact and fetch it back verified	docs
awdit	An append-only audit trail whose gaps are DETECTABLE	docs
awbac	Role-based access control that fails closed and explains itself	docs
awiam	Who is this caller? A directory and session store that fails honestly	docs
awtunnel	Reach a service that has no public address	docs
awnest	Prove there is a human before you let them into the nest	docs
awnboard	A front gate you can put in front of anything, and hand someone the key to	docs
awnix	A Linux you can hand to an agent — immutable base, capabilities included	docs
awrecover	Labelled snapshots with an all-or-nothing restore	docs
awkno	The man page for the Aither World — every brick, stack and law, offline	docs
awrelay	Portable agent messaging — findings, alerts, coordination	docs
awmail	Give an agent an email address — send, and actually receive	docs
awnet	The agentic web — agents host a mesh, and agents join one	docs
awfind	A portable search client — query, results, ranking	docs
awbrowse	A portable browser client — navigate, console, network, DOM, screenshot	docs
awknowledge	How to run a coding agent so the result survives — the laws, with evidence	docs
aitherkvcache	Near-optimal KV cache quantization for LLM inference — sub-byte compression	docs
AitherZero	PowerShell 7+ automation framework — numbered, self-describing scripts	docs
AitherConnect	Browser extension — federated AI search, page context, and the Living OS overlay	docs
awreason	A portable reasoning client — sessions, phases, thoughts, and the chain that produced the answer	docs
awrecurse	Answer a question over a context far larger than the window — recursively, with the trace kept	docs
awprism	Turn a failure into ranked hypotheses — and say what would confirm each one	docs
awrepl	A REPL an agent can actually use — state that survives between turns	docs
awresearch (you are here)	Ask a research question, get a cited report you can check	docs
awpredict	Predict what your environment does next, and how surprised you were	docs
awkno	The man page for the Aither World — every brick, stack and law, offline	docs

<script src="aither-constellation.js"></script>