Aegis Latent Core
Self-hosted evidence gateway that signs and commits a verifiable record for every LLM call before the response is released.
Pick Aegis when the question you must answer is not "what did the model say" but "who could have altered the record." An offline-checkable MMR inclusion proof beats a vendor log export, and committing blocked traffic into the same chain closes the gap most audit trails leave open. The release discipline is unusually honest: v5.0.1 documented its Sigstore tag and cosign-verified GHCR images with 31 assets rather than paper over gaps. Go in eyes open — you run it, it is fail-closed, and the project claims no certification. If you want a vendor on the hook for uptime, buy managed guardrail SaaS instead.
Verified 4h ago · liveness 71/100 · cite: rightaichoice.com/tools/aegis-latent-core
- Security and compliance teams in regulated industries that must prove AI usage to an auditor
- Platform teams self-hosting an LLM gateway to keep custody of evidence and keys
- Organizations with EU AI Act Art. 12, HIPAA, SEC 17a-4 or MiFID II record obligations
- Engineering teams running multi-provider LLM apps that need policy plus verifiable evidence
- Teams that want a managed SaaS guardrail with a vendor operating it for them
- Low-traffic projects with no audit requirement beyond ordinary application logs
- Groups without the ops capacity to run and tune a fail-closed gateway
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Aegis Latent Core if you want a managed guardrail with an uptime SLO and a vendor on the hook, rather than running a fail-closed gateway and holding the signing keys yourself.
Because the gateway is fail-closed, a degraded signer, limiter or durable store takes your LLM traffic down with it — budget redundancy for those components, not just the app tier
The gateway is free at $0 under Apache-2.0 — the real cost is engineering time to run it plus durable storage, which puts it in range for a small platform team that would find a managed guardrail SaaS seat bill hard to justify. The COMMERCIAL.md license is custom-priced, so distributed-product vendors should price that negotiation before they commit architecture.
In short
Aegis Latent Core — Self-hosted evidence gateway that signs and commits a verifiable record for every LLM call before the response is released. Best for Security and compliance teams in regulated industries that must prove AI usage to an auditor, Platform teams self-hosting an LLM gateway to keep custody of evidence and keys, Organizations with EU AI Act Art. 12, HIPAA, SEC 17a-4 or MiFID II record obligations. Free to use.
What's new in Aegis Latent Core
Checked todayAcross the latest 2 updates: 1 launch and 1 changelog entry.
v5.0.1 — signed tag, 31 release assets, cosign-verified GHCR images
Release published across all surfaces except PyPI aegis-latent-core; the Sigstore-signed tag passes gitsign verify-tag, 15 files in SHA256SUMS re-hash to their digests, and the SDK is byte-identical on PyPI and npm.
v5.0.0 published on releases and GHCR
Major release published on the same surfaces as v5.0.1; the project notes there is no 4.2.0 version.
What people actually say about Aegis Latent Core — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
26 mentions across 3 sources (YouTube, Product Hunt, Lemmy) · researched Aug 28, 2026.
Average across the 3 sources that answered — each source counts once, not each post.
- +Addresses a clear regulatory need for AI usage accountability
- +Emphasizes tamper-proof audit logs for legal and compliance use
- +Self-hosted, offering control and data privacy for enterprises
- +Supports OpenAI-compatible routes and native Anthropic Messages traffic
- +Positioned for regulated industries like finance, healthcare, and government
- −No independent reviews or user experiences to validate claims
- −Product Hunt launch got zero upvotes, indicating minimal interest
- −No details on pricing, integrations, or deployment requirements
- −All available information comes from the founder, creating bias
- −Unknown reliability at scale, especially as a self-hosted gateway
- • Potential costs for self-hosting infrastructure, maintenance, and security patches
- • Potential costs for custom integrations or professional services
Viability Score
How well maintained and how widely used is Aegis Latent Core? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Commits signed, hash-linked evidence before the response reaches your caller
- Append-only Merkle Mountain Range evidence log
- O(log n) inclusion proofs verified offline by an independent third party
- verify_integrity() tamper detection on record read
- Fail-closed admission: no signer, limiter or durable store means no service
- WAF rules applied to LLM traffic, with blocks committed to the evidence chain
- Per-user and per-session rate limiting with quota breaches recorded
- Egress control on outbound model calls
- Admission control covering authentication, scope and request bounds
- Request redaction on the governed path plus session management
- OpenAI-compatible API surface with a configurable upstream endpoint
- Bounded streaming semantics with a declared commit point
- Durable storage with WAL replay
- Dashboard for browsing committed evidence
- Python and TypeScript SDKs for offline proof verification
About Aegis Latent Core
Aegis Latent Core is an Apache-2.0 gateway you deploy between your application and one or more LLM providers. It admits, bounds, redacts and forwards governed calls, and a response is not released until the evidence path has reached its declared commit point. Records go into an append-only Merkle Mountain Range log, so an independent third party can check an O(log n) inclusion proof offline with the published Python or TypeScript SDK — verification is the product, not a report. Refused traffic lands in the same chain: WAF blocks and quota breaches are committed rather than discarded. Controls around that path include admission control for auth, scope and bounds, WAF rules written for LLM traffic, per-user and per-session rate limiting, egress control on outbound model calls, request redaction, session management, and durable storage with WAL replay, plus a dashboard for browsing committed records. An OpenAI-compatible surface means upstream is just a configured endpoint. The latest published release read here is v5.0.2 (2026-10-01) on GitHub, PyPI aegis-latent-core and PyPI/npm aegis-latent-sdk; the Sigstore-signed tag, 31 release assets, cosign-verified GHCR images and SHA256SUMS readback belong to v5.0.1 (2026-09-24). The audience is narrow on purpose: AI platform teams, security engineers, regulated operators, incident responders and auditors who need a self-hosted control plane rather than a hosted dashboard or a compliance badge.
Behind the Verdict
Aegis Latent Core is best understood as a control plane with an evidence spine, not a logging add-on. Every governed call passes admission control (authentication, scope, request bounds), WAF rules written for LLM traffic, per-user and per-session rate limits, and egress control before it is forwarded to the upstream endpoint; the request is redacted on the way through, and the response is held until the declared commit point is reached. That ordering is the whole design: a fail-closed admission path means no signer, no distributed limiter or no durable store means no service, which is exactly what you want in a regulated request path and exactly what will page you at 2am if the store is unhealthy. The differentiator is verifiability. Records land in an append-only Merkle Mountain Range log, and the Python and TypeScript SDKs let an independent third party check an O(log n) inclusion proof offline — no access to your database, no trust in your vendor. verify_integrity() flags tampering on read, and refused traffic (WAF blocks, quota breaches) is committed to the same chain instead of being dropped, so the record of what you blocked is as checkable as the record of what you allowed. The OpenAI-compatible surface keeps provider choice open: upstream is a configured endpoint, so multi-provider apps do not get locked in by the gateway itself. Where it fits: EU AI Act Art. 12 record-keeping, HIPAA audit trails, SEC 17a-4 and MiFID II obligations, and any incident-response workflow where you must reconstruct decisions after the fact. Where it does not: teams without ops capacity to run and tune a fail-closed gateway, low-traffic projects whose ordinary application logs satisfy everyone, and buyers who need a certificate or a published uptime SLO on a procurement form today. Read the release history carefully before you plan an upgrade. v5.0.0 and v5.0.1 (2026-09-16 and 2026-09-24) shipped with signed tags, SHA256SUMS and cosign-verified GHCR images; the current v5.0.2 (2026-10-01) is published on GitHub, PyPI aegis-latent-core and PyPI/npm aegis-latent-sdk, and the README makes no GHCR verification claim for that cycle — the stronger Sigstore and cosign readbacks belong to v5.0.1. Budget real time for standing up the gateway and dashboard containers plus durable storage before your first governed call, and get legal sign-off on AGPL-3.0 early if the gateway sits in a distributed product.
Researching Aegis Latent Core? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Aegis Latent Core actually fits — and what changes day-one when you adopt it.
Stand the gateway up between the internal app fleet and two model providers, route all calls through one OpenAI-compatible endpoint, and set admission, WAF and per-user rate limits on day one.
Outcome: Every governed call — allowed or refused — lands in the MMR log with a commit point, so the audit team can pull an inclusion proof instead of asking engineering for a log export.
Investigate a suspected prompt-injection attempt by reading WAF-blocked requests and their policy decisions out of the same append-only chain as successful calls.
Outcome: You can show exactly what was blocked and when, without trusting a separate log pipeline that may have dropped the records.
Take a record and its inclusion proof, run the published Python or TypeScript SDK offline, and call verify_integrity() with no access to the operator's systems.
Outcome: A reproducible verification result that does not depend on the operator's honesty or the vendor's uptime.
Use Cases
- Prove AI usage to an auditor with an offline-checkable inclusion proof instead of a vendor-hosted log export
- Enforce AI usage policy across internal teams routing through one governed endpoint
- Commit evidence of refused or quota-breached requests so blocked traffic is auditable
- Satisfy AI governance record obligations under EU AI Act Art. 12, HIPAA, SEC 17a-4 or MiFID II
- Run multi-provider LLM applications behind an OpenAI-compatible gateway without provider lock-in
- Self-host a gateway where the compliance team, not a vendor, holds the signing keys
Limitations
- The project makes no certification and no SLO claim, and availability is your deployment's problem: fail-closed means no signer, no distributed limiter or no durable store means no service.
- Tuning, upgrades and on-call are yours.
- The core is AGPL-3.0, which some legal teams will not accept on a service in the request path; the repository's COMMERCIAL.md describes a commercial license alternative.
- Read release artifacts before you upgrade: v5.0.1 (2026-09-24) carried a Sigstore-signed tag, 31 release assets, cosign-verified GHCR images and re-hashed SHA256SUMS, while the current v5.0.2 (2026-10-01) is published on GitHub, PyPI aegis-latent-core and PyPI/npm aegis-latent-sdk with no GHCR verification claim in the README for this cycle.
- Budget time for gateway and dashboard containers plus durable storage before your first governed call.
as of 2026-10-18
Verification history
We have re-verified Aegis Latent Core 5 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-checked, vendor evidence unchanged
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Aegis Latent Core tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Open Source (Apache-2.0)
$0
Ideal for
Platform or security teams who will self-host the gateway and hold their own signing keys and durable evidence storage
What this tier adds
Starting tier at $0: full gateway with fail-closed policy, WAF and egress controls, signed MMR evidence log, offline Python/TypeScript verification and the records dashboard
Commercial License
Custom
Ideal for
Vendors shipping a distributed product that cannot accept AGPL-style copyleft obligations on a gateway in the request path
What this tier adds
Adds an alternative to copyleft terms for distributed products, with terms set out in COMMERCIAL.md and pricing arranged with the maintainer
Where the pricing makes sense
The company stage and team size where Aegis Latent Core's pricing actually pencils out — and where peers do it cheaper.
The gateway is free at $0 under Apache-2.0 — the real cost is engineering time to run it plus durable storage, which puts it in range for a small platform team that would find a managed guardrail SaaS seat bill hard to justify. The COMMERCIAL.md license is custom-priced, so distributed-product vendors should price that negotiation before they commit architecture.
Setup time & first value
How long it actually takes to get something useful out of Aegis Latent Core — broken out by persona, not the marketing-page minute.
Platform team: expect a working day or more before the first governed call — gateway and dashboard containers plus durable storage and a signer, then policy and upstream endpoint. A single engineer evaluating it: an afternoon to get a proof verifying through the SDK. Third-party auditor: minutes, since the Python or TypeScript verifier runs offline.
Switching to or from Aegis Latent Core
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From plain application logs: route the same traffic through the gateway's OpenAI-compatible surface and let committed records replace log-line evidence
- →From a managed guardrail SaaS: point your existing OpenAI-compatible client at the self-hosted endpoint and re-express vendor policies as admission, WAF and rate-limit rules
- →From an earlier Aegis release: move to v5.0.2 from the GitHub Release, PyPI aegis-latent-core or PyPI/npm aegis-latent-sdk, and check the SHA256SUMS and signed tag history if you need cosign-grade artifact verification
- ↗To a managed guardrail SaaS: swap the OpenAI-compatible base URL back to the vendor endpoint, accepting that evidence custody and signing keys leave your control
- ↗To ordinary log shipping: export records before decommissioning, since inclusion proofs and fail-closed commits have no equivalent in a log pipeline
Resources & Guides
- Resourcegithub.com
DEPLOYMENT GUIDE · Aegis Latent Core
Helpful link from github.com
- Resourcegithub.com
README · Aegis Latent Core
Helpful link from github.com
- Documentationgithub.com
Docs · Aegis Latent Core
Full product docs from github.com
- Resourcegithub.com
SUPPORT · Aegis Latent Core
Helpful link from github.com
- Resourcegithub.com
SECURITY · Aegis Latent Core
Helpful link from github.com
- Examplesgithub.com
Examples · Aegis Latent Core
Working sample projects from github.com
- API Referencegithub.com
Sdk · Aegis Latent Core
Methods, params, types from github.com
- Resourcegithub.com
Llms · Aegis Latent Core
Helpful link from github.com
Tutorials & Learning
YouTube returned 6 videos for “Aegis Latent Core”, and we withheld 6: 6 did not mention Aegis Latent Core. We are showing none, because we could not prove any of them are about Aegis Latent Core.
Official links
Featured Head-to-Head Comparisons
Aegis Latent Core vs Mindgard
If your priority is actively attacking and defending AI systems—especially agents—Mindgard is the clear choice: it automates red teaming, maps attack surfaces, and has a track record of public disclosures. Choose Aegis Latent Core only if your primary need is passive governance and audit trails for LLM traffic, not offensive testing.
Aegis Latent Core vs Arize Phoenix
If your priority is enforced compliance and verifiable audit evidence for enterprise LLM traffic, Aegis Latent Core is the safer bet. But for AI engineers actively building and debugging agents, Arize Phoenix is the clear winner — it’s open-source, self-hostable, and packed with tracing and evaluation tools. Pick based on whether you need a governance gate or a development workbench.
Aegis Latent Core vs Persefoni
Choose Persefoni if your pain point is mandatory climate reporting—it's a mature, AI-enhanced carbon accounting platform with clear regulatory alignment and a freemium entry. Choose Aegis Latent Core if you need to govern and audit every LLM interaction inside your enterprise; it's the missing piece for AI compliance, but you'll need to talk to sales and it lacks the breadth of Persefoni's feature set. Both serve different masters.
Popular in AI Governance & Guardrails
Mindgard
Mindgard automates AI red teaming — agent-native reconnaissance, exploit validation, and continuous AI security testing for production AI
Poolside AI
Poolside builds open-weight agentic coding models — Laguna XS 2.1 (33B) and Laguna S 2.1 (118B) — for code that stays inside your security
Olas Network
Co-own autonomous AI agents on-chain: stake OLAS, run agents that trade, hire each other and settle payments.
Frequently Asked Questions
Categories
Best-of guides
Topics
Used Aegis Latent Core? Help shape our editorial sentiment research.