v0.20.0

Changelog

• 32ed2f6 Merge remote-tracking branch 'marbor-new/main' into new-name
• f954dba branding: capitalize Marbor + improve banner styling
• d7ca35a build(deps): bump lucide-react from 1.27.0 to 1.33.0 in /ui (#81)
• 589c519 build(deps): bump modernc.org/sqlite from 1.56.0 to 1.57.0 (#77)
• 8f69d5e build(deps-dev): bump @vitejs/plugin-react from 6.0.5 to 6.1.0 in /ui (#80)
• 188eac8 build(deps-dev): bump eslint-plugin-react-refresh in /ui (#79)
• 8165c2e build(deps-dev): bump vite from 8.2.1 to 8.2.2 in /ui (#78)
• ec6c270 chore(release): v0.20.0
• 1d5a8cb chore(rename): complete Phase 1 hard-cutover codebase rename to marbor
• 3182923 chore: complete remaining mesh -> marbor rename (UI, internal, CLI, configs, docs)
• d1e0851 chore: complete residual mesh -> marbor rename
• 9edd2fd chore: regenerate marbor.demo.db for marbor_agent schema rename
• ed6b455 ci(pages): add timeout to deploy job
• 7e96f10 docs(bench/ui): purge dead BENCH-RUNBOOK.md pointers, rename-era MESH_URL vars, real ui README
• 74580b4 docs(readme): centered hero header with marbor branding
• 7ca5b5b docs(runtime): document why MLX is indistinguishable from llama.cpp
• 1cee94b feat(activity): unified fleet activity feed with kind taxonomy and CLI parity
• 784849e feat(admin): expose agent-stale flag on the nodes API
• b82fb12 feat(dashboard): pair Saved-vs-Cloud with Fleet Capacity above the fold
• 56b7b74 feat(dashboard): restructure around fleet health and capacity
• 2a4949e feat(demo): Cloud Providers card fully interactive on /demo/
• 5482666 feat(ui): collapsible sidebar nav with persisted desktop toggle
• 04f2400 fix(admin): bound HF config.json architecture fields against int64 overflow
• 0df1529 fix(admin): check crypto/rand errors, serialize schedule writes, encode JSON properly, audit login/reset
• fdc6722 fix(admin): close login role/status enumeration oracle
• 5e5c043 fix(admin): decode settings body outside the shared config lock
• 09e4fc0 fix(admin): distinguish unreadable system memory from a genuine 0 MB reading
• b5ecda2 fix(admin): eliminate settings-PATCH map aliasing and lost-update races
• ddcfc35 fix(admin): extend P131 overflow guard to the GGUF/Ollama arch input path
• d44ee05 fix(admin): gofmt SystemInfo struct field alignment
• 09fd943 fix(admin): make benchmark job ID collision-proof
• 3a5f207 fix(admin): stop fabricating "all RAM free" when MemAvailable is exhausted
• 38e27b6 fix(admin): stop mislabeling agent/api VRAM readings as nvidia-smi
• 4d70a12 fix(admin): surface ErrUnloadUnsupported as a clear 4xx, not an opaque 502
• 1bf9b83 fix(admin): validate HF repoID, default runtime to node's declared value, cap config cache
• 0c12ac8 fix(agent): add HTTP server timeouts to close slowloris exposure
• 03b82f1 fix(agent): add missing argv terminator to unload command
• 811b62f fix(agent): cap request body size on model pull/delete/unload handlers
• 7f9af74 fix(agent): correct process-liveness detection across three drivers
• 6c6f9f1 fix(agent): explicit --enroll always takes precedence over an existing token
• 98f4226 fix(agent): reject a service Config with only one of CertPath/KeyPath set
• 4aee270 fix(agent): reject leading-dash model names and add argv terminators
• c764ffe fix(agent): reject pid<=0 in readPIDFile
• a2c9145 fix(agent): remove ineffective liveProc PID-reuse guard
• c87addc fix(agent): require admin scope for arbitrary runtime commands
• f6117c6 fix(agent): retain live process handle to close a PID-reuse window
• ae8a760 fix(agent): retry transient runtime-version failures, recover refresh panics
• b9dc518 fix(ansible): register-gpus.yml asserted pre-rename mesh_url var, breaking every run
• 57b231a fix(ansible): remove duplicate register key breaking retry/health-poll
• dbd774d fix(ansible/docs): agent enrollment sent the runtime port as the AGENT's port
• eda6cb9 fix(audit): log a rate-limited warning when AppendAuditLog fails
• b88d6cb fix(auth): hold the lock across Refund's limiter lookup-plus-refund
• 5e6317f fix(auth): make Reload's key-map swap atomic
• b0531f6 fix(auth): parse date-form expires_at in local time, not UTC
• e25ed77 fix(bench): normalize --target trailing slash, drop racy drain goroutine, fix rounding
• 1999334 fix(bench): validate N, check prerequisites, fix scratch-file cleanup and error paths
• cf0efa1 fix(bench): verify actual VRAM eviction before firing a cold TTFT sample
• 02f8a19 fix(ci): add timeout-minutes to the four jobs missing one
• 8f601be fix(ci): fail release job loudly on Pages-deploy dispatch error
• 12b2cd6 fix(ci): register QEMU emulation for arm64 docker release builds
• 01c335d fix(ci): scope pages.yml permissions to the job that needs them
• f713bdc fix(cli): classify 404/405/409/422 as user errors on read-only GET paths
• 89a8e8b fix(cli): fail non-zero when the completion script write fails
• 31b1798 fix(cli): gate 503 passthrough on the health-check path only
• ea5f1b8 fix(cli): local session read failure is a user error, not a server error
• 2d425b3 fix(cli): repair bash/zsh completion after a positional argument, and its syntax entirely
• c5b1462 fix(cli): surface terminal-restore errors, restore echo on SIGINT
• bbffdfe fix(cli): write the session file atomically via temp-then-rename
• 5cdc188 fix(compose): bind Grafana to loopback by default in the monitoring overlay
• b721ad9 fix(config): avoid octal misinterpretation in numeric-IP SSRF guard
• 3a1c27c fix(config): close SSRF gap on cloud providers/webhook, reject duplicate key tokens
• 23c6f85 fix(config): close SSRF guard bypass via integer-encoded hosts
• 15aeaa9 fix(config): extend port-collision check to admin, validate session-affinity TTL, dedupe repeat wildcard-bind warning
• 4326a6b fix(demo): hash the seeded demo admin session token
• b54b584 fix(demo): honor stream:false, bind listeners before printing the banner
• 2069739 fix(demo): inline the demo database so the standalone compose file is self-contained
• 06603ae fix(demo): opt demotraffic into its own non-localhost guard for the compose network
• c40fb06 fix(demo): strip seeded mlx-node row, bind admin/metrics to loopback
• cfe831a fix(demotraffic): distinguish a truncated stream from a genuine 0 tokens
• e7a566a fix(demotraffic): mark synthetic traffic and require opt-in for non-localhost
• 6ea8160 fix(docs): aggregate mismatched labels in the cloud-fallback-rate alert
• 252616f fix(docs): correct grammar left over from mesh -> marbor rename
• 420ced2 fix(gen-docs): mark required flags in group-page FLAGS section
• d1ee4d8 fix(infra): CI hardening, image pinning, arm64 docker, install.ps1 TLS (P365-P374)
• 49f40fd fix(infra): installer sudo -E exit handling for agent role
• 3a817e4 fix(infra): installer/gate/smoke hardening, docs-site URL scheme allowlist (P375-P386)
• 58b4ac7 fix(install): default to a per-user path for non-elevated Windows installs
• c93e8c7 fix(install): disable crash-looping systemd unit before nohup fallback
• 67abce5 fix(install): quote systemd unit paths so a space doesn't break parsing
• f47e776...

v0.19.3

Changelog

• 6b7665c Merge fix/node-agent-scheme-decoupling: decouple Node Agent TLS scheme from runtime URL
• 0c1ca6b chore(demo-db): regenerate mesh.demo.db for new node_agent.scheme column
• 782a4b8 chore(deps): bump framer-motion from 12.42.2 to 13.1.0 in /ui (#75)
• eb15319 chore(deps): bump golang.org/x/crypto from 0.54.0 to 0.55.0 (#71)
• 8a065d2 chore(deps-dev): bump @types/react-dom from 19.2.3 to 19.2.4 in /ui (#72)
• ca6f1e7 chore(deps-dev): bump typescript-eslint from 8.65.0 to 8.67.0 in /ui (#74)
• 3b2cfe2 chore(release): v0.19.3
• 57c46dd feat(model-advisor): context-length feasibility advice (P71)
• e37fbe6 fix(model-advisor): harden P71 context-feasibility correctness
• a3880c4 fix(nodeagent): update stale ollama-mesh-agent invocation strings in agent.go
• 8620c7b fix(router,admin): decouple Node Agent TLS scheme from runtime URL
• 1a85258 fix(ui): add confirm dialog to Regenerate Token (R10)
• f60ca49 fix(ui): address code-review findings on Agent modal restructuring
• 3e8eeed fix(ui): align Agent Port, HTTPS, and Save Connection on one row
• 6366680 fix(ui): box the Agent Connection/TLS Certificate sections in Manage Agent modal
• 74aa7dd fix(ui): don't show plain green Enabled for an unusable HTTPS Agent
• 080231e fix(ui): move Regenerate Token / Disable Agent next to Health Check
• 91e0007 fix(ui): remove redundant status line above the Agent Connection card
• 2d9fb7c fix(ui): separate Runtime and Node Agent config on GPU Nodes page (P90)
• 94ab50f refactor(model-advisor): cache P71 HF/node lookups, dedupe HF fetch, simplify recommendation

v0.19.2

Changelog

• 51b6d70 Merge remote-tracking branch 'origin/main'
• 1315f3c chore(release): v0.19.2
• f589b8a feat(ansible): first-party fleet registration playbook
• 0c9c983 feat(ansible): split fleet enrollment into two independent playbooks
• 0c02937 feat(nodeagent): split Node Agent into a separate ollama-mesh-agent binary
• fffbc5c fix(admin): parse hourly analytics retention keys as UTC
• 9ef57bb fix(admin): reject benchmark runs when proxy port is unconfigured
• 48aafe7 fix(admin): restore per-key policy fields dropped on config reload
• 20a057e fix(admin): use writeJSONError for handleModelRepo's missing-id error
• d389cd4 fix(admin): use writeJSONError instead of hand-built JSON error strings
• 79e98a1 fix(admin,ui): update stale ollama-mesh agent invocation strings
• 09c96df fix(bench): true median calculation, preserve evict/parse error detail
• 4044027 fix(demo): poll for proxy readiness instead of a fixed 500ms sleep
• 9ffbbec fix(demotraffic): stop silently clamping REQUEST_COUNT, flag missing eval_count
• e91a2fd fix(docs): correct fleet-enrollment guide auth model, wire copy buttons on doc site
• a6c4080 fix(loadtest): compute RPS from actual elapsed time, not nominal duration
• bb515c9 fix(mocknode): deterministic OpenAI token-family matching, honor stream:false
• bc827c9 fix(router): POST /admin/nodes upserts by name instead of duplicating live capacity
• ebe3cb9 fix(ttft): drain response body synchronously instead of racing Close()

v0.19.1

Changelog

• 9b24a4e chore(deps): bump modernc.org/sqlite from 1.54.0 to 1.56.0 (#66)
• 9ea8ea6 chore(deps): bump react-router-dom from 7.18.1 to 7.18.2 in /ui (#58)
• 590ce01 chore(deps): bump recharts from 3.10.0 to 3.10.1 in /ui (#57)
• 1afe73d chore(deps-dev): bump @types/node from 24.13.2 to 26.2.0 in /ui (#59)
• 6d892d6 chore(deps-dev): bump eslint from 10.7.0 to 10.8.1 in /ui (#60)
• 291db20 chore(deps-dev): bump vite from 8.1.5 to 8.2.1 in /ui (#61)
• 9b4ad4e chore(release): v0.19.1
• 241856e feat(cli): add "ollama-mesh uninstall" and error on unknown subcommands
• fca1102 feat(ui): click-to-copy on truncated Request Log values (Request ID, Key, Model)
• 0754a45 fix(cli): registry-driven CLI dispatch, strict arg validation, help/docs generation (P83+)
• 83dfd7a fix(install): add Node Agent install dir to PATH on Windows
• 2c60aa0 fix(nodeagent): reject unrecognized agent subcommand instead of silently running foreground
• 519d732 fix(nodeagent): stop agent service status from showing a stale TLS fingerprint
• b9c6fb2 fix(ui): give Request Log table an explicit width instead of implicit auto
• c3f2987 fix(ui): make Request Log Key badge hug its text, not its column width
• 06c48a1 fix(ui): make Request Log table fill its card with calibrated proportions
• 057298e fix(ui): replace click-to-copy with in-place scroll for long Request Log values
• 66a6f4c fix(ui): stop Request Log columns ballooning on wide viewports
• 121e370 fix(ui): stop Request Log table overflowing on long key/id/model values
• 1b8b770 fix(ui): stop Request Log's card wrapper stretching past the table
• 3f56b4d fix: strip internal roadmap phase tag (P24) from public --help/usage text
• 376d7fc test(admin): prove tls_fingerprint null is a no-op, not a clear (P24)

v0.19.0

Changelog

• 364a28a Merge branch 'claude/deprecate-plaintext-token-cli-858c38'
• 91f4370 Merge pull request #54 from Anirudhx7/claude/model-advisor-pull-runtimes-827f51
• 5cea93c Merge pull request #55 from Anirudhx7/claude/mlx-demo-pull-capability
• f3f6447 Merge pull request #62 from Anirudhx7/claude/ollama-mesh-release-consolidation-1bfb43
• 0892dbe Merge pull request #63 from Anirudhx7/feature/p66-local-only-spill
• 8663c17 Merge pull request #64 from Anirudhx7/feature/p67-local-degradation-chain
• 2b22bbb Merge pull request #65 from Anirudhx7/feature/p67-per-key-degradation-policy
• 8c54602 Merge pull request #67 from Anirudhx7/p54-node-agent-token-scoping
• 33e94ae Merge pull request #69 from Anirudhx7/feature/p24-node-agent-tls
• 61dd2d6 Merge pull request #70 from Anirudhx7/feature/p41-routing-explainability
• 0422875 chore(db): regenerate mesh.demo.db for allow_local_degradation column
• 653d08e chore(demo): regenerate mesh.demo.db for P66 schema (spill_counters, local_only)
• 4d4fa92 chore(release): v0.19.0
• d71fd24 chore: regenerate mesh.demo.db for P41 routing_reason/routing_detail columns
• bc32b56 chore: regenerate mesh.demo.db for node_agent.scope column (P54)
• 808ad05 chore: regenerate mesh.demo.db for node_overrides.gpu_indices column
• 707dbb3 chore: regenerate mesh.demo.db for node_overrides.tls_fingerprint (P24)
• f24644c docs(admin): correct P70's CLI-parity claim, note namespaced-tag ambiguity inline
• 10a0f7c docs(config): fix stale header references in local-degradation doc comments
• 52aaff2 feat(admin): TLS fingerprint validation and tls-probe endpoint (P24)
• 4aa2df7 feat(admin): expose local_only + spill counters through the admin API
• b465240 feat(auth): add per-key local_only policy field
• 597ae4d feat(auth): replace client header with per-key allow_local_degradation policy
• b22c264 feat(bench): add SQLite write-path load sweep tool, publish capacity numbers (P53)
• 349fbbb feat(cli): add login/logout/whoami with a persisted session file
• 2319b14 feat(cli): add mesh key set-local-only and mesh spill
• 8f62757 feat(cli): add nodes confirm-tls subcommand (P24)
• ff47718 feat(cli): close CLI parity gaps for models and runtime Group 2 actions
• 54cdbb8 feat(nodeagent): TLS certificate generation and TLS listener (P24)
• 0fc143a feat(proxy): fail closed on local_only keys, wire spill accounting
• f9d8dea feat(router): TLS fingerprint pinning and dial-time verification (P24)
• 40c8bd0 feat(router): hard per-node in-flight cap that sheds instantly (P64) (#68)
• 721fb7d feat(router,admin,cli,ui): per-request routing explainability (P41)
• 4951169 feat(routing): local-first degradation chain before cloud egress (P67)
• 6d5313b feat(store): add TLS fingerprint column with downgrade-safe upsert (P24)
• fbe0d35 feat(store): add spill_counters table and per-key local_only column
• 85239ad feat(ui): TLS fingerprint pinning UI (P24)
• 2bd9e93 feat(ui): add MB/GB unit selector for VRAM total override
• d53b69a feat(ui): local_only toggle on API keys, Cloud Spill section on Analytics
• c854db1 feat(ui): optional paste-to-verify for TLS fingerprint pinning
• ca685cf feat(ui): searchable dropdown for the fallback-chain alternates field
• e211494 feat(ui): searchable model-name dropdown for Settings config cards
• a9d27af fix(admin): close HF search-layer runtime asymmetry for vllm/tgi/mlx (P77)
• dbc9889 fix(admin): declare per-node GPU scope to fix Advisor double-counting and disclose unknown GPU counts (P75 Gaps B/C/D)
• ff9a687 fix(admin): exclude mmproj vision-projector files from GGUF variant list
• 86b9ebd fix(admin): exclude sharded multi-part GGUF files from variant list
• 81e71e3 fix(admin): extract shared disk-telemetry guard, document P73 in CHANGELOG
• 4c853c7 fix(admin): fix multi-GPU VRAM fit for handleModelFit and largest-basis free calc
• 348172c fix(admin): gate Model Advisor catalog fit and pull on runtime/format compatibility (P70)
• a0b1ec1 fix(admin): hard-block model pulls with unknown size on thin disk headroom (P73)
• 5397aa8 fix(admin): patch declared-GPU-scope edge cases found in P75 code review
• b7d9d61 fix(admin): serialize handlePatchNode's TLS validate-to-mutate transaction
• 9c5bd65 fix(admin): size Model Advisor fit against all local GPUs, not GPU 0 only (P69)
• 40302f2 fix(admin): stop excluding vision-language models from vllm/tgi/mlx search (P77 follow-up)
• fbac14a fix(admin): validate resulting TLS pin on URL moves
• 0712e6d fix(admin,ui): make multi-fetch search concurrent/partial-tolerant, fix pipeline_tag badge display (P77 audit)
• c966076 fix(admin,ui): show Model Advisor Pull button for agent-capable non-Ollama nodes
• 05d6363 fix(agent): agent service start/stop --help ignored args, ran the real action
• dc60a77 fix(agent): agent service status --help also ignored args, ran the real check
• c6e7c32 fix(bench): cap loadtest concurrency, correct README env-var claim
• 0000f71 fix(ci): rebuild mesh.demo.db from a clean main checkout
• 8067e95 fix(ci): regenerate demo admin password hash to actually match "admin"
• 0e41dcd fix(ci): update smoke test CLI check for removed --token flag
• a89f4ef fix(cli): address code-review findings on login/logout/whoami
• 86e0645 fix(cli): bench/agent --help now writes to stdout instead of stderr
• 22e075f fix(cli): kill env-secret leak via --help, and 3 more auth findings
• 4ccc088 fix(cli): reject unknown models actions instead of silently listing all models
• da55604 fix(cli): tab-align help tables, add real per-subcommand action help, fix --help exit code
• 6e0f4cc fix(cli): update top-level --help banner for new commands and stale tagline
• 4e02f4b fix(config): reject reserved local/blocked cloud provider names
• 646097e fix(demo): regenerate mesh.demo.db against the fixed admin password hash
• 9fe27cd fix(nodeagent): add per-action token scoping (P54)
• 26b92ec fix(nodeagent): collapse repeated CLI progress lines to one error message
• faaa81a fix(nodeagent): fail closed on unrecognized dotted token scope prefix
• efb7626 fix(nodeagent): fix flaky disk-stats test; expose docker_deployed flag
• 50d23c8 fix(nodeagent,admin): check container's real disk space before a Docker pull
• 30db5fd fix(proxy): harden cloud fallback for /api/embeddings and /api/embed (P78 Phase 1)
• 42f239b fix(proxy): stop truncating unbounded single-JSON responses and clamping unavailable tokens to zero (P80)
• f9f1d31 fix(proxy,admin): preserve real HTTP status through request logging (P82)
• 81145db fix(router): add hard model-identity eligibility filter (P79)
• 884e61e fix(router): preserve/re-derive NodeState.Host across UpdateNodeURL
• 2a3ab01 fix(routing): enforce single-hop local degradation, per-key allow-list, and affinity safety
• bda49dc fix(security): close Node Agent token leak paths (P68)
• 8291b02 fix(security): deprecate plaintext --token CLI args for Node Agent enrollment
• 12b86b7 fix(security): harden UserSession.Token with json:"-"
• ddd5c14 fix(security): remove plaintext --token CLI flag from admin CLI
• 5b772c0 fix(ui): add chevron toggle to the fallback-chain alternates combobox
• b8f806e...

v0.18.8

Changelog

• 19dfbe8 fix(nodeagent,admin): route model delete/unload through docker exec on containerized runtimes
• f77e88a fix(nodeagent,admin,ui): route model pull through docker exec on containerized runtimes

v0.18.7

Changelog

• 88c91b8 chore(demo): regenerate mesh.demo.db to match current schema
• bb75935 chore(demo-db): regenerate mesh.demo.db for runtime_nodes.host column
• 31f6724 chore(deps-dev): bump typescript-eslint from 8.62.0 to 8.65.0 in /ui (#52)
• 8d753ad docs(changelog): record audit-driven placement/persistence fixes
• f57ba69 feat(router,admin,ui): surface warmup suppression state on the dashboard
• a008766 fix(admin,audit): move hot-path audit-log and local-stat writes off request path
• 6c88a31 fix(nodeagent,router,admin): make Node Agent host-scoped, not runtime-scoped
• 82f2489 fix(router): detect model-name/digest collisions in warm-residency scoring
• 871b5b8 fix(router): drain must block warmup pings and scheduled WarmModels
• b928534 fix(router): interpret free VRAM per-runtime, not universally as free bytes
• fa91606 fix(router): narrow VRAM reservation race in findBestByScore
• 4e01d8d fix(router): penalize stale telemetry in placement scoring
• cb33159 fix(router): predictive warmup engine must respect unload suppression
• 4b3a8d4 fix(router,store): fix node-agent host-scoping gaps from 6c88a31
• 85d3738 fix(router,store): persist session affinity across restart
• ad36139 fix: pause console on Windows fatal exit instead of closing instantly

v0.18.6

Changelog

• 1433db0 fix(admin): gate runtime start/stop/restart/logs on agent reachability, not runtime health
• 6f8bd50 fix(nodeagent): exclude agent's own service unit from runtime discovery
• 46614dc fix(ui): require confirmation before clearing a node's control driver

v0.18.5

Changelog

• f3d0f18 Merge branch 'feature/p58-runtime-logs' into main
• 0c01d28 Merge remote-tracking branch 'origin/main' into feature/p58-runtime-logs
• 7f5aa9d chore(deps): bump github.com/prometheus/client_golang (#47)
• 8c99319 chore(deps): bump lucide-react from 1.24.0 to 1.27.0 in /ui (#50)
• a43ba07 chore(deps-dev): bump @vitejs/plugin-react from 5.2.0 to 6.0.4 in /ui (#48)
• e12c2b7 chore(deps-dev): bump globals from 16.5.0 to 17.8.0 in /ui (#49)
• c25dfd8 chore(deps-dev): bump vite from 8.1.4 to 8.1.5 in /ui (#51)
• 04dfadf feat(admin): add runtime.logs Admin API dispatch (P58)
• 538f9d9 feat(cli): add mesh runtime logs command (P58)
• d16ee37 feat(nodeagent): add runtime.logs ControlDriver capability (P58)
• 21fc349 feat(ui): add View Logs button to Runtime Control panel (P58)
• 8a68aed fix(auth): recreate rate limiter on Reload and clean up key_counters on delete
• da704cf fix(nodeagent): fix Windows service reinstall race and install.ps1 window close

v0.18.4

Changelog

• dace3d3 fix(nodeagent): handle Windows SCM service control dispatcher via sys/windows/svc to fix error 1053