This release fixes issues for SREs watching stability and regressions.
✓ No known CVEs patched in this version
Topics
+5 more
ReleasePort's take
Light signalgraphify v0.7.15 fixes Ollama integration bugs where invalid context window settings could exhaust VRAM, and improves CLI help and version output.
Why it matters: Prevents VRAM exhaustion when GRAPHIFY_OLLAMA_NUM_CTX exceeds available memory; adds warnings for configuration mismatch. Improves CLI help and version output. Routine upgrade for Ollama users.
Summary
AI summaryFixed help flag handling to stop execution and added explicit warnings for token‑budget mismatches.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Bugfix | Medium |
Help flags in any position now stop execution and print usage Help flags in any position now stop execution and print usage Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Bugfix | Medium |
--version, -v, and graphify version now print installed version --version, -v, and graphify version now print installed version Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Bugfix | Medium |
Invalid GRAPHIFY_OLLAMA_NUM_CTX no longer defaults to hardcoded value exhausting VRAM Invalid GRAPHIFY_OLLAMA_NUM_CTX no longer defaults to hardcoded value exhausting VRAM Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Bugfix | Medium |
Warns when GRAPHIFY_OLLAMA_NUM_CTX smaller than chunk size, suggests correction Warns when GRAPHIFY_OLLAMA_NUM_CTX smaller than chunk size, suggests correction Source: llm_adapter@2026-05-21 Confidence: high |
— |
Full changelog
Bug Fixes
- #821 Universal
-h/--help/-?guard: any help flag in any position now stops execution and prints usage — previouslygraphify cursor install --helpsilently installed into Cursor, andgraphify benchmark --helpcrashed withFileNotFoundError: '--help' - #818
--version,-v, andgraphify versionnow print the installed version and exit — previously fell through to "unknown command" - #820
GRAPHIFY_OLLAMA_NUM_CTX=<invalid>no longer falls back to hardcoded131072(which exhausted VRAM on constrained cards) — now falls through to auto-derived value with a warning - #820 When
GRAPHIFY_OLLAMA_NUM_CTXis pinned smaller than the estimated chunk size, graphify now warns explicitly that Ollama will silently truncate the prompt and suggests a corrected--token-budget
Upgrade
pip install -U graphifyy
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About graphify
AI coding assistant skill (Claude Code, Codex, OpenCode, Cursor, Gemini CLI, OpenClaw, Factory Droid, Trae). Turn any folder of code, docs, papers, images, videos, or YouTube links into a queryable knowledge graph
Beta — feedback welcome: [email protected]