Sevana Voice Quality
Server Details
Voice quality analysis: PVQA and AQuA MOS for WAV files, VoIP call analysis for PCAPs
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 19 tools
Upload, job-control, and analysis tools are clearly separated, and the pcap report tools generally have distinct retrieval purposes. There is some overlap among pcap_summary, pcap_full_report, pcap_network_stats, and pcap_get_stream, but the descriptions make the granularity differences mostly understandable.
All names use snake_case and helpful namespace prefixes such as pcap_, pvqa_, job_, and aqua_. The action placement is slightly inconsistent (pcap_get_call vs. pcap_summary vs. pcap_analyze), but the overall pattern remains predictable.
Nineteen tools is on the heavy side for this domain, largely because of the seven pcap reporting tools. Each has a plausible role, but the surface is expansive enough to feel borderline rather than tightly scoped.
The server covers upload lifecycle, intrusive and non-intrusive WAV analysis, pcap VoIP analysis, job control, and detailed SIP/RTP/DTMF reporting. Minor gaps such as listing uploaded files or exporting a consolidated report exist, but core workflows are well covered.
Available Tools
19 toolsaqua_compareAQuA: compare with a referenceARead-onlyInspect
Intrusive voice quality analysis (Sevana AQuA, wideband): compares a degraded/test WAV with its reference WAV and returns MOS, percent similarity and a fault analysis explaining the loss of quality. Upload both files first.
Inputs: test_upload_id (required), reference_upload_id (required), channel_test, channel_reference. Returns: mos, faults, percent, test_upload_id, +1 more
create_uploadCreate an upload URLAInspect
Start uploading a WAV recording (kind "wav") or a packet capture (kind "pcap"). Returns an upload_id and a one-time upload_url; send the file as the body of an HTTP PUT to that URL (for example: curl -T recording.wav '<upload_url>'), then pass the upload_id to an analysis tool. The URL works once and expires after 15 minutes; the file is deleted after an hour. Use upload_inline instead if you can only pass tool arguments.
Inputs: kind, filename, size_bytes. Returns: method, max_bytes, upload_id, expires_at, +3 more
delete_uploadDelete an uploadBIdempotentInspect
Delete an uploaded file now instead of waiting for it to expire.
Inputs: upload_id (required). Returns: deleted
job_cancelCancel a jobBIdempotentInspect
Cancel a queued or running analysis job.
Inputs: job_id (required). Returns: next, error, stage, job_id, +7 more
job_statusJob statusARead-onlyInspect
State of an analysis job (queued, running with stage and percent, done, failed, cancelled). Waits up to wait_seconds for it to finish; a finished PCAP job includes its summary.
Inputs: job_id (required), wait_seconds. Returns: next, error, stage, job_id, +7 more
list_jobsList jobsARead-onlyInspect
Your analysis jobs from the last hour, newest first.
Returns: jobs
pcap_analyzePCAP: analyze a captureARead-onlyInspect
Start a VoIP quality analysis of an uploaded .pcap/.pcapng capture (upload it with kind "pcap"). Reconstructs SIP calls and RTP streams and scores every stream (Sevana MOS from the decoded speech, network MOS, loss, jitter). Returns a job_id; the call waits up to wait_seconds and, if the analysis finished by then, includes the summary. Otherwise poll job_status. Only analysis reports are returned, never decoded audio. Reports include SIP messages and DTMF digits as captured; they are kept for one hour.
Inputs: upload_id (required), statistics, wait_seconds. Returns: next, error, stage, job_id, +7 more
pcap_dtmfPCAP: DTMFARead-onlyInspect
DTMF key presses found in a finished PCAP analysis, with digits and timing: RTP telephone-events (RFC 2833/4733) per stream and SIP INFO (application/dtmf-relay, application/dtmf).
Inputs: job_id (required), call_id, stream_id. Returns: count, digits, events
pcap_full_reportPCAP: full report (paged)ARead-onlyInspect
The complete analysis report, one section at a time: "streams" or "sip_calls" (full objects, paged with cursor/limit) or "counters" (packet counters, SIP statistics, aggregate statistics). Contains no audio.
Inputs: job_id (required), section (required), limit, cursor. Returns: total, section, streams, software, +6 more
pcap_get_callPCAP: one SIP callARead-onlyInspect
One SIP call of a finished PCAP analysis: peers, user agents, response codes, timing, the SIP message ladder (method or status line per message), health warnings and the call's streams.
Inputs: job_id (required), call_id (required). Returns: health, ladder, peer_a, peer_b, +15 more
pcap_get_streamPCAP: one RTP streamBRead-onlyInspect
Full detail of one RTP stream: quality (Sevana and network MOS, R-factor, impairments, echo), loss gaps, payload types, DTMF events. include_timeline adds PVQA's per-interval (0.68 s) detector table.
Inputs: job_id (required), stream_id (required), include_timeline. Returns: stream, summary
pcap_network_statsPCAP: network statisticsARead-onlyInspect
Network view of a finished PCAP analysis: packet and SIP counters, per-stream loss, jitter, delay and network MOS, and the aggregate statistics (correlations of MOS with network features) when computed.
Inputs: job_id (required). Returns: streams, statistics, sip_statistics, packet_counters
pcap_sip_messagesPCAP: SIP messagesARead-onlyInspect
SIP messages of a finished PCAP analysis, exactly as captured (headers and SDP bodies), in time order. Filter by call_id; paged with cursor/limit.
Inputs: job_id (required), limit, cursor, call_id. Returns: total, messages, next_cursor
pcap_summaryPCAP: summaryARead-onlyInspect
Overview of a finished PCAP analysis: packet totals, SIP calls with their health warnings, and streams ordered worst Sevana MOS first, each with codec, MOS, loss, jitter and impairments.
Inputs: job_id (required), max_calls, max_streams. Returns: calls, job_id, omitted, packets, +7 more
pvqa_analyzePVQA: voice quality of a recordingARead-onlyInspect
Non-intrusive voice quality analysis (Sevana PVQA) of an uploaded WAV file: predicted MOS (1-5) and the impairments found (low SNR, noise, dead air, clipping, clicks, echo, packet-loss artifacts, ...). No reference recording is needed. report_type "extended" adds the per-interval detector table.
Inputs: upload_id (required), channel, report_type. Returns: mos, channel, rfactor, timeline, +5 more
pvqa_detectorPVQA: raw detector valuesARead-onlyInspect
Raw values of one PVQA detector over time for an uploaded WAV file. Detectors: SNR, DeadAir-00, DeadAir-01, Click, VADClip, AmpClip, DynClip, Echo, Noise, Crushed, DTMF.
Inputs: detector (required), upload_id (required), channel. Returns: rows, names, detector, upload_id
pvqa_timelinePVQA: impairment timelineBRead-onlyInspect
Impairment timeline of an uploaded WAV file: which impairments PVQA detects at which time.
Inputs: upload_id (required), channel, resolution_ms. Returns: rows, channel, columns, upload_id, +1 more
server_infoServer, limits and quotaARead-onlyInspect
Analyzer versions, your tier (free or API key), your limits and how many analyses you have left.
Returns: role, tier, limits, server, +3 more
upload_inlineUpload a small file inlineAInspect
Upload a WAV or pcap file as base64 in the tool arguments (2 MB at most after decoding). Returns an upload_id for the analysis tools. Larger files: use create_upload.
Inputs: content_base64 (required), kind, filename. Returns: wav, filename, upload_id, expires_at, +1 more
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Measure voice/VoIP path quality -> estimated MOS + live network metrics
Simulation, evaluation and monitoring for voice agents.
Pronunciation assessment, phoneme scoring, speaker voice ID, audio transcription, speech synthesis.
Pronunciation scoring, speech-to-text, and text-to-speech for language learning
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceEnables agents to compare a processed audio file against its reference and report how much perceptual quality was lost, using PEAQ ODG, ViSQOL MOS-LQO and an optional licensed POLQA binary, plus waveform, loudness and Bark-band spectral metrics. Runs headless over stdio, installs itself with uv, and returns rankings across codecs or bitrates.GPL 3.0
- FlicenseNot gradedqualityDmaintenanceMeasure the network path quality to a voice/VoIP destination and get an estimated MOS (ITU-T G.107 E-model) plus live RTT, jitter, and packet loss. Agent-native: REST + MCP, free tier then pay-per-call via x402.-
- FlicenseNot gradedqualityDmaintenanceEnables comprehensive video quality analysis including metadata extraction, GOP structure analysis, quality metrics comparison (PSNR, SSIM, VMAF), artifact detection, and transcoding effect assessment through FFmpeg-based tools.6-
- AlicenseAqualityCmaintenanceEnables local analysis of PCM WAV files to produce bounded audio observations such as activity segments, clipping indicators, duration, sample rate, peak, and RMS level without uploading recordings.1MIT
Glama MCP Gateway
Add one secure layer between your agents and this server.