The autonomous workspace
for deep research & inquiry.
Launch Aeris for macOSDownload Aeris
Y CombinatorFinding website design example
Today’s research workflows are broken. Traditional AI assistants hallucinate citations, lose context across dense multi-page documents, and fail when tasks require deep primary-source verification and multi-step synthesis.
Aeris is an autonomous workspace built for deep inquiry. It orchestrates specialized research agents in parallel across scientific repositories, financial filings, and internal databases—delivering verified conclusions with transparent provenance.
Any scientific paper, dataset, or filing,
Aeris synthesizes with precision.
Unlike shallow LLM summaries, Aeris executes deep multi-hop reasoning across thousands of primary sources, extracting verified data points, validating claims, and structuring publication-ready findings with full citation transparency.
The SOTA research intelligence engine.
Aeris ranked #1 on multi-hop scientific reasoning and citation accuracy benchmarks, achieving 99.4% factual recall across complex multidisciplinary corpora. Learn more
Memory that retains your research context.
Aeris maps your project hypotheses, annotated literature, and synthesis history into persistent local knowledge graphs. All memory stays strictly local on your device and is never shared with third parties. Learn more
Private vault that
*************.
Proprietary research shouldn’t leak to public models.
Aeris isolates internal patents, confidential trial notes, and raw lab data with end-to-end local encryption.
Data stays invisible to cloud LLMs Citations and document vectors are queried locally through zero-knowledge embedding layers.
Human verification at the edge Critical synthesis outputs and external queries always await your explicit confirmation.
Full provenance audit log Every retrieved conclusion links back to the exact paragraph, table, and data source.
Aeris Vault
The zero-knowledge research vault built for teams.
Hardware-backed E2E encryption
Your research library is fully encrypted on-device and protected by Apple Silicon Secure Enclave.
Scoped permissions and source provenance
Grant synthesis agents access only to selected project folders. Verify what was accessed and when.
******* enough for confidential research.
Aeris processes your research locally, enforces zero-data-retention, and guarantees your proprietary discoveries never train third-party models.
Local by default Notes, literature embeddings, and synthesis artifacts stay strictly on your device.
Encrypted by your key Everything is secured at rest using AES-256 and Apple Secure Enclave encryption.
Sandboxed runtime Aeris isolates memory and file access with kernel sandboxing to protect your intellectual property.
Bring your own models Connect local Ollama / LM Studio models or route securely via your own Anthropic and OpenAI enterprise keys.