groundlens
OfficialGroundLens is a read-only MCP server that verifies AI answers the passages they were written from, plus execution traces and record logs.
verify_answer — Checks an answer against its sources (id/text pairs,
{id,text,provenance}dicts, or bare strings) with no model in the loop; acceptspolicy(built-in name, path, or YAML),question,scope,risk_class, andlocale; returns a PASS/REVIEW/FAIL decision with reasons, evidence, regulatory mapping, and the signed record with its content hash.verify_run — Verifies an MCP execution trace (JSON-RPC messages as JSON Lines) under an execution policy, given
run_id,system, and optionalstarted_at/system_version; returns the gate (ALLOW/REVIEW/DENY), any breaches, and the signed run record.verify_records — Verifies a JSON Lines log of answer- or run-records offline, checking every hash, link, and signature; returns
{ok, verified, kind}and fails if any record or link was altered.

Guardrails for LLM answers
Quickstart
# pip install groundlens
import groundlens as gl
verifier = gl.Verifier(sink=gl.sinks.jsonl("records/"))
record = verifier.check(answer="Revenue was €42.1 million.", passages=["Revenue reached €41.2 million."])
print(record.decision, record.reasons)FAIL ['c2: groundlens.numeric contradicted: €42.1 million not in sources; nearest 41200000 EUR in ctx-0 (relative distance 0.021)']Runs locally: no account, no API key, no model, no network. The signed record is in records/ before check returns.
Related MCP server: Sentry MCP
Integration
The same check, wired into your stack:
from langfuse import get_client
verifier = gl.Verifier(sink=gl.sinks.jsonl("records/"), observers=[gl.integrations.langfuse(get_client())])Call check inside the observation that produced the answer, and its trace gets:
Score
groundlens.decision: categorical,PASS,REVIEWorFAIL, on the check and on its trace. Filter traces by it, chart it, export it.Observation
groundlens.check: typeguardrail. The reasons are its status message, one per line. Its level follows Langfuse's own scale:PASSisDEFAULT,REVIEWisWARNING,FAILisERROR.Metadata of the check: decision, reason codes (
groundlens.scope.contradicted), policy, and the id and hash of the signed record behind it.Judge: a
generationobservation with the model name, next to your own model calls.Spans: one per verifier, plus the store, each with its duration. Log View shows them in milliseconds.
SDK: Python v3 and v4. Tested on Langfuse Cloud and self-hosted Langfuse 4.58.
verifier = gl.Verifier(sink=gl.sinks.jsonl("records/"), observers=[gl.integrations.otel()])Spans and metrics go through the providers your application already configured: Tempo, Prometheus, or any OpenTelemetry backend.
Metrics:
groundlens_decisions_total{decision, policy},groundlens_reasons_total{reason, decision},groundlens_check_duration_seconds{stage}with stagetotal,engine,judge,sign,store.Spans:
groundlens.checkwithgroundlens.decision,groundlens.reasons, record id and hash; children per verifier, the judge (gen_ai.*) and the store.Dashboard: the one in the image.
gl.integrations.grafana_dashboard()returns the path of its JSON; import it in Dashboards → New → Import.PromQL, share of answers that failed:
sum(rate(groundlens_decisions_total{decision="FAIL"}[15m])) / sum(rate(groundlens_decisions_total[15m]))Alert when more than 5 % fail:
- alert: GroundLensFailRateHigh expr: sum(rate(groundlens_decisions_total{decision="FAIL"}[15m])) / sum(rate(groundlens_decisions_total[15m])) > 0.05 for: 10mTry it:
docker run -p 3000:3000 -p 4318:4318 grafana/otel-lgtm, thenpython examples/observability/credit_copilot.py grafana --minutes 5. Tested on Grafana 12 with Tempo and Prometheus.
Your assistant checks its own draft before it answers. Three read-only tools over stdio: verify_answer, verify_run, verify_records. Needs uv.
Claude Code (plugin: the MCP server plus a skill that tells Claude when to check):
claude plugin marketplace add groundlens-dev/groundlens claude plugin install groundlens@groundlensCodex:
codex mcp add groundlens -- uvx --from "groundlens[mcp]" groundlens mcpAny MCP client (Claude Desktop, Windsurf, Zed...):
{ "mcpServers": { "groundlens": { "command": "uvx", "args": ["--from", "groundlens[mcp]", "groundlens", "mcp"] } } }With the package installed (
pip install "groundlens[mcp]"),"command": "groundlens-mcp"is enough.Nothing to install: the Hugging Face Space is also a remote MCP server, with
verify_answer, athttps://groundlens-groundlens.hf.space/gradio_api/mcp/.
A skill on ClawHub, the skill registry of OpenClaw. Before your assistant sends an answer with figures taken from your documents, it runs groundlens verify on its draft:
PASS: the answer goes out.
FAIL: it fixes the figure from the source and checks again.
REVIEW: the answer goes out, saying which figure it could not confirm.
Runs on your machine, with no API key and no second model. The skill is in plugins/groundlens/skills/groundlens, the same one the Claude Code plugin uses. OpenClaw can also use the MCP server directly, with openclaw mcp add.
check() returns only after the signed record is stored. If it cannot be stored, check() raises and no decision comes back. Each sink keeps a chain: every record carries the hash of the one before it.
Sink | What you pass | How it stays append-only |
| a directory or a | one JSON line per record, flushed to disk and locked on every write. For development and single machines |
| your S3 client and a bucket created with Object Lock | one object per record, written only if absent; |
| an | one blob per record, never overwritten; give the container a time-based immutability policy |
| any DB-API connection | one row per record; the table is created if missing, and its keys stop a position from being written twice or a chain from forking |
import boto3
verifier = gl.Verifier(
signer=gl.signers.aws_kms(boto3.client("kms"), key_id), # the key never leaves KMS
sink=gl.sinks.s3_object_lock(boto3.client("s3"), "audit-records", retain_days=3650),
)Several processes: give each its own chain (chain="worker-3") and verify them together with groundlens record verify. GroundLens depends on none of these SDKs: you pass the client you already use. Postgres, S3 Object Lock, AWS KMS, Azure Blob and PKCS#11 are tested in CI against real software (Postgres, MinIO, LocalStack, Azurite, SoftHSM); Azure Key Vault has no emulator and is experimental.
LLM-as-a-judge
The exact checks catch a figure that is wrong, miscalculated, from another company or of another kind than asked. They cannot catch a real figure taken from the wrong row of the right table, or a sentence the passages never say. For that you connect a model that reads: the judge. GroundLens ships no model; you pass the client you already use.
from openai import OpenAI
verifier = gl.Verifier(
"credit_desk.yaml", # a policy that allows a judge, below
judge=gl.judges.openai_compatible(OpenAI(), "gpt-4.1"),
sink=gl.sinks.jsonl("records/"),
)Adapter | Client |
|
|
|
|
|
|
The shipped policies forbid a judge, so the policy that uses one says so:
id: credit_desk
version: 1.0.0
extends: eu_ai_act_high_risk_v1
verifiers:
forbidden: [] # allow llm_judge.*
judge:
required: true # no judge, or a judge that fails: REVIEW
contradicted: REVIEW # or FAIL
unsupported: REVIEWThe judge can escalate, never pass. It labels each sentence supported, unsupported, contradicted or uncertain. supported resolves nothing on its own, so a judge that is wrong, or that a passage has talked into agreeing, cannot let through an answer the exact checks fail. A timeout, an error or a malformed reply is REVIEW.
You see exactly what it is sent. One prompt per answer, from a fixed template (groundlens.judge/1): what PASS means for you, the scope, the question, the answer and every passage with its provenance. Passages sit between markers they cannot contain, so text inside them is data, not instructions. groundlens judge show-template prints the exact prompt before anything is sent anywhere.
The record keeps the trail. The template id, the hash of the prompt, the model the provider says answered and the hash of its reply, with each sentence's label, the passages it cited and its one-line reason. In Langfuse the judge appears as a generation, next to your own model calls.
Architecture
GroundLens splits verification into five parts that never mix: claims, verifiers, evidence, policy and record.
The answer is cut into claims: each figure and each sentence, with its character offsets.
Verifiers examine each claim against the passages and return evidence (
supported,contradicted,unsupported), never a verdict. The shipped ones are exact: numeric (exact fractions, locales, units, recomputed sums and ratios), scope, intent, and your rules. The judge, if you connect one, is the model you already use, asked through a fixed template.The evidence graph keeps what each verifier said about each claim, and where they disagree.
The policy, a YAML file you own, turns evidence into a decision:
PASS,REVIEWorFAIL, with reasons.The record holds the input hash, the evidence, the policy hash and the decision, chained to the previous record and signed. It is stored before
checkreturns; observers are told after.
The engine is Rust, shipped as a Python wheel (abi3, Python 3.10+), with a standalone glv binary that checks records without Python. It runs in your process: no server, no network, no runtime dependencies. From Rust, cargo add groundlens; for the binary, cargo install groundlens-cli.
Parameters
Parameter of | Options |
|
|
|
|
|
|
|
|
|
|
| how numbers are read and compared |
Parameter of | |
| required; passages as strings or |
| enables the intent check: the answer gives the kind and unit asked for |
| what the request is about ( |
| a class your policy can send to human review whatever the checks say |
Policy
The checks produce evidence. The policy decides what that evidence means for you: what fails, what goes to a person, and what PASS allows. It is a short YAML file that you own and version with your code. The same answer under two policies can get two different decisions, and both are correct; that difference is your risk appetite, written down.
Two policies ship with GroundLens:
Policy | For |
| any application. A contradicted figure or a figure from outside the scope fails; an answer of the wrong kind or unit goes to review. No model decides. |
| systems under Annex III of the EU AI Act. The same rules, with each decision mapped to Art. 15(1) accuracy, Art. 14(4)(a) human oversight and Art. 12(1) record keeping. |
Write your own policy
Extend a shipped policy and write only what differs:
id: credit_desk
version: 1.2.0
extends: eu_ai_act_high_risk_v1
pass_means: "may be shown to the analyst as a summary of the filing"
locale: es-ES # how the documents write numbers: 1.234,56
rounding: half_even # how a recomputed figure may round
scope:
required: [entity] # every request must name the borrower
missing_scope: REVIEW # or FAIL; never PASS
risk_classes:
default: credit_information
mandatory_review: [credit_limit_change] # always seen by a personverifier = gl.Verifier("policies/credit_desk.yaml")
verifier.check(answer=a, passages=p, scope={"entity": "Talleres Norte"}, risk_class="credit_limit_change")Key | What it sets |
| a shipped policy to start from. Maps merge key by key, lists replace. Your policy must set its own |
| what PASS allows in your application, in one sentence. A verifier's PASS is not a decision to publish; this says which one it is. |
| how figures are read and compared: |
| the keys every request must carry, and what a missing one gets. |
| classes that always go to review. A class the policy does not list is an error, so a typo cannot route around review. |
| which checks are |
| which verifiers can fail an answer, which send it to review, and what unresolved or conflicting claims get. |
| whether a judge is required and what its labels may do. See LLM-as-a-judge. |
| which control and article each decision evidences, in your own control ids. |
Unknown keys are errors, not ignored: a misspelt rule never silently does nothing.
Failure is not a PASS
Situation | Decision |
a figure contradicts the passages, under the rounding rule | FAIL |
a figure comes only from a passage outside the scope | FAIL |
a required scope key is missing | REVIEW (or FAIL, if the policy says so) |
a risk class that requires review | REVIEW, whatever the checks say |
a check could not run, or the judge timed out or failed | REVIEW |
the record cannot be stored | an exception, and no decision |
The final decision is the worst outcome of all checks.
Check your policy
groundlens policy lint policies/credit_desk.yaml # errors before you deploy
groundlens policy show --resolved policies/credit_desk.yaml # every default filled inThe hash of the resolved policy goes into every record it decides, so each decision can be traced to the exact rules that produced it.
Before you change a policy,you have to measure it:
groundlens measureestimates the share of wrong answers among those that passed, from a small sample your team labels without seeing the verdicts, with a 95 % confidence interval, per risk class, and the cost per decision.groundlens replayruns past traffic under the new policy and lists the decisions that would change.
Both reports are signed. Full reference: Writing policies.
Security
GroundLens sits on the path between a model and a decision, and its records are meant to be evidence. Both make it a target: an attacker may try to get a wrong answer through, or to alter the trail afterwards.
Threat | Control |
A wrong answer passes | Exact checks with no model in the default path; errors and timeouts are |
Instructions hidden in passages steer the judge | Passages are delimited data in a fixed, versioned template; the judge can only escalate; adversarial tests |
A record is edited, deleted or reordered | SHA-256 content hash, hash chain, signature; |
The signing key leaks | Keys stay in AWS KMS, Azure Key Vault or a PKCS#11 HSM (ECDSA P-256); GroundLens never holds them |
Two writers fork a chain | Conditional writes in every sink; a lost race is retried, never overwritten |
Data leaves through the verifier | No network code in the engine ( |
Hostile input exhausts the parser | Bounded numeral grammar against regular-expression DoS; fuzzing of every parser |
A tampered release | Pinned and hashed CI dependencies, SLSA Build Level 3 provenance, signed release binaries, OpenSSF Scorecard |
Report vulnerabilities privately through GitHub security advisories. See SECURITY.md.
Contributing
If GroundLens got one of your answers wrong, that case is the most valuable contribution there is. Checks, connectors and documentation are welcome too; CONTRIBUTING.md explains how to start.
Open source
Apache-2.0, all of it: engine, checks, policies, connectors, command line and MCP server. Verification evidence has to be checkable by anyone, without our software or our permission, and that only holds if the verifier is open.
Contact
You can contact the developer at javier@groundlens.dev
1. GroundLens 6.0.5, default policy, no judge. Financial answers: 189 real answers by Claude Haiku from a RAG over FinQA 10-K pages (e5 embeddings, top 4 passages), on multi-step questions never used while developing GroundLens. Each answer labelled twice by a language model (Claude Opus 5.5, claude-opus-5-5) against fixed written criteria, blind to GroundLens, with a third pass on the 3 disagreements; labels were frozen before GroundLens ran. 22 answers were wrong (1 in 9): 17 got FAIL or REVIEW (77 %, 95 % CI 57–90 %). Of the 129 answers that passed, 5 were wrong (3.9 %, 95 % CI 1.7–8.8 %). The price: 60 of 189 answers (32 %) went to FAIL or REVIEW, including 38 of 110 correct ones. Median 10.6 ms per check in a fresh process. Public claims: LLM-AggreFact test split (29,320 claim–document pairs from 11 sets, human labels), never used while developing GroundLens. Of the 1,616 unsupported claims that contain a number, 796 got FAIL or REVIEW (49 %, 95 % CI 47–52 %); among the claims with a number that passed, 16.3 % were unsupported, against 25.4 % before the check. Claims without a number all passed: GroundLens caught none of their errors. Median 4.0 ms per check. All results, including the sets where it does not help, in BENCHMARK.md.
Copyright Groundlens 2026
Available Tools
3 toolsverify_answerA
Check an answer against the passages it was written from, before you send it.
Every figure must be in the passages or follow from figures that are; no model
is used. Returns PASS, REVIEW or FAIL with the reasons, and the signed record.
sources: (id, text) pairs, {"id","text","provenance"} dicts, or bare strings.
policy: a built-in name (e.g. "eu_ai_act_high_risk_v1"), a path, or YAML.
scope: what the request is about, e.g. {"entity": "UPS"}; a figure taken
from a source whose provenance is another entity fails.
risk_class: the request's risk class, when the policy defines them.
Returns the decision (PASS/REVIEW/FAIL), the evidence, the regulatory
mapping and the record with its content hash.
Inputs: answer (required), sources (required), scope, locale, +3 more. Returns: object
verify_recordsA
Verify a log of records offline: every hash, every link, every signature.
records: the JSON Lines text of an answer-record or run-record log.
Returns {"ok", "verified", "kind"}; fails if any record or link was altered.
Inputs: records (required). Returns: object
verify_runA
Verify an MCP execution trace under an execution policy and return the run record.
trace: the MCP session as JSON-RPC messages (JSON Lines).
policy: the execution policy, as YAML/JSON text or a path.
Returns the gate (ALLOW/REVIEW/DENY), any breaches, and the signed run record.
Inputs: trace (required), policy (required), run_id (required), system (required), +2 more. Returns: object
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 3 tools
Each tool takes a clearly different input shape: a JSON-RPC session trace, answer passages, and a JSON Lines record log. There is mild overlap because verify_records can validate run records that verify_run also produces/validates, which could cause a brief hesitation, but the described inputs keep them separable.
All three tools follow a strict verb_noun pattern (verify_run, verify_answer, verify_records), with the same verb used consistently for the same conceptual action.
Three tools is on the lean side but each covers a distinct verification target (run trace, answer grounding, record log), so nothing feels redundant or padded.
The core verification lifecycle (verify runs, verify answers, verify logs offline) is covered, and records are produced by the verify calls themselves. A minor gap is that no tool enumerates available built-in policies (e.g. 'eu_ai_act_high_risk_v1'), though agents can work around this by passing policy text or a path.
Maintenance
Related MCP Connectors
Fact-checks generated content against your sources of truth showing what to trust, change, & verify.
Check AI-written drafts against your source documents, with the passage behind every verdict.
Real-time fact-check, citation verification, and source-freshness for AI agents.
DraftCheck: flags AI-writing patterns in prose with fix hints. Deterministic linter, not a detector.
Related MCP Servers
AlicenseBqualityFmaintenanceAn MCP server that provides a comprehensive interface to Semgrep, enabling users to scan code for security vulnerabilities, create custom rules, and analyze scan results through the Model Context Protocol.61,203 PyPI689MIT
Sentry MCPofficial
AlicenseAqualityBmaintenanceA remote Model Context Protocol server acting as middleware to the Sentry API, allowing AI assistants like Claude to access Sentry data and functionality through natural language interfaces.74971Functional Source , Version 1.1, ALv2 Future- Apache 2.0

Arkheia Hallucinationofficial
AlicenseNot gradedqualityCmaintenanceDetect fabrication and hallucination in any LLM output. Score responses from GPT-4o, Claude, Gemini, Llama and 30+ models. Free tier included.MIT