Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.
C++
launch: shorter model-list wording (too big, SSD N GB) so it fits 110 columns; docs and previews for the classic-console look and several cards
Niko1221fa726b4c9ce1
diff --git a/docs/TERMINAL_LAUNCH.md b/docs/TERMINAL_LAUNCH.mdindex fe62b120e..ffb22d40c 100644--- a/docs/TERMINAL_LAUNCH.md+++ b/docs/TERMINAL_LAUNCH.md@@ -46,7 +46,7 @@ server's `/metrics`, the same data as the browser Monitor), the other five views(`c` copy, `w` write it for you after a backup, `h` check it: one prompt and one tool call sent the way that app sends them). `q` asks: keep serving in the background, or stop(stopping unloads the model and frees the graphics memory).-In the classic Windows console (Consolas) the monitor draws ASCII sparklines and square corners instead of block glyphs the font does not have (`STRATA_GLYPHS=unicode|ascii` forces either).+In the classic Windows console (the Windows 10 default window) the monitor first switches that window's font to Cascadia Mono when it is installed (the console font API: no admin, nothing saved, gone with the window) and draws the full look; otherwise it draws a solid-block look: █ bars with a thin ▄ track, ▄ / █ sparklines, square boxes and the 16 palette colours (`•` instead of `●` when the font is not Consolas). It always shows one line: "For the full look, use Windows Terminal (built into Windows 11)". Windows Terminal, Linux and macOS are never touched. `STRATA_GLYPHS=unicode|ascii|solid` forces a look. With two or more graphics cards the Overview tiles show the busiest card (load), all cards (memory), the hottest card and the total power, a "Cards" strip lists each card (up to 4, then "press 5" for the Hardware view), and the CPU tile says "2 CPUs · 64 cores" from the server's new `sockets` figure.## Without the screens (flags)diff --git a/docs/terminal-previews/launch-monitor-overview.txt b/docs/terminal-previews/launch-monitor-overview.txtindex bf53d31db..4ecc490ed 100644--- a/docs/terminal-previews/launch-monitor-overview.txt+++ b/docs/terminal-previews/launch-monitor-overview.txt@@ -19,8 +19,8 @@└────────────────────────┘ └────────────────────────┘ └────────────────────────┘ └────────────────────────┘Context fill 36K / 128K 28% Recent requests############................................ TIME STATUS PROMPT REUSED OUTPUT TOK/S DURATION- Experts in VRAM 1,644 · 3.2 GB 20:34:50 Done 17,183 0 19,511 50.5 429.8 s- ##############.............................. 20:34:46 Done 626 0 11 26.8 4.0 s+ Experts in VRAM 1,644 · 3.2 GB 23:53:40 Done 17,183 0 19,511 50.5 429.8 s+ ##############.............................. 23:53:36 Done 626 0 11 26.8 4.0 sSystem RAM 58.5 / 63 GB#########################################...diff --git a/docs/terminal-previews/launch-overview-ascii.txt b/docs/terminal-previews/launch-overview-ascii.txtindex 599c7245f..bc1a44fe1 100644--- a/docs/terminal-previews/launch-overview-ascii.txt+++ b/docs/terminal-previews/launch-overview-ascii.txt@@ -18,8 +18,8 @@ Model state Idle Reading [Generating] Queued Error└────────────────────────┘ └────────────────────────┘ └────────────────────────┘ └────────────────────────┘Context fill 36K / 128K 28% Recent requests############................................ TIME STATUS PROMPT REUSED OUTPUT TOK/S DURATION-Experts in VRAM 1,644 · 3.2 GB 20:34:53 Done 17,183 0 19,511 50.5 429.8 s-##############.............................. 20:34:49 Done 626 0 11 26.8 4.0 s+Experts in VRAM 1,644 · 3.2 GB 23:53:42 Done 17,183 0 19,511 50.5 429.8 s+##############.............................. 23:53:38 Done 626 0 11 26.8 4.0 sSystem RAM 58.5 / 63 GB#########################################...diff --git a/docs/terminal-previews/launch-start.txt b/docs/terminal-previews/launch-start.txtindex bf7713d1b..9fbf4f379 100644--- a/docs/terminal-previews/launch-start.txt+++ b/docs/terminal-previews/launch-start.txt@@ -1,7 +1,7 @@# strata launch preview: start screen with saved presets# (text of the screen as drawn at the size named in tests/terminal/launch_previews.py)- STRATA Launch RTX 5070 12 GB · 64 GB RAM · 52 GB free+ STRATA Launch RTX 5070 12 GB · 64 GB RAM · 47 GB free────────────────────────────────────────────────────────────────────────────────────────────────────────────Start from a saved preset@@ -31,6 +31,6 @@────────────────────────────────────────────────────────────────────────────────────────────────────────────- This computer: RTX 5070 12 GB · 64 GB RAM · 52 GB free. Preset: model, context, parallel requests, tuned+ This computer: RTX 5070 12 GB · 64 GB RAM · 47 GB free. Preset: model, context, parallel requests, tunedsettings, access. Shared with the Strata app.↑↓ choose Enter launch e edit preset d delete ? help q quitdiff --git a/docs/terminal-previews/launch-step1-model.txt b/docs/terminal-previews/launch-step1-model.txtindex 8652b9c24..53d6b2aa9 100644--- a/docs/terminal-previews/launch-step1-model.txt+++ b/docs/terminal-previews/launch-step1-model.txt@@ -3,7 +3,7 @@STRATA Set up Step 1 of 9: Model────────────────────────────────────────────────────────────────────────────────────────────────────────────- Which model? This computer: RTX 5070 12 GB · 64 GB RAM · 52 GB free+ Which model? This computer: RTX 5070 12 GB · 64 GB RAM · 47 GB freeModel Download Needs (RAM + VRAM) On this computerQwen3.8-Flash-Next Q2_0 66 GB 35 GB fitsQwen3.8-Flash-Next IQ2_XS 68 GB 36 GB fits@@ -14,15 +14,15 @@Swift 1.5 IQ3_S 84 GB 51 GB fitsQwen3.8-Flash-Next Coder IQ1_M 58 GB 24 GB fitsQwen3.8-Flash-Next (Unsloth) UD-IQ4_XS 94 GB 60 GB fits, part read from SSD- Qwen3.8-Flash-Next (Unsloth) UD-Q4_K_XL 111 GB 78 GB fits · uses SSD 16 GB- Qwen3.8-Flash-Next (Unsloth) UD-Q5_K_XL 158 GB 99 GB tight · uses SSD 37 GB- Qwen3.8-Flash-Next (Unsloth) UD-Q6_K_XL 169 GB 110 GB tight · uses SSD 48 GB- Qwen3.8-Flash-Next (Unsloth) Q8_0 188 GB 129 GB tight · uses SSD 67 GB+ Qwen3.8-Flash-Next (Unsloth) UD-Q4_K_XL 111 GB 78 GB fits · SSD 16 GB+ Qwen3.8-Flash-Next (Unsloth) UD-Q5_K_XL 158 GB 99 GB tight · SSD 37 GB+ Qwen3.8-Flash-Next (Unsloth) UD-Q6_K_XL 169 GB 110 GB tight · SSD 48 GB+ Qwen3.8-Flash-Next (Unsloth) Q8_0 188 GB 129 GB tight · SSD 67 GBQwen3.8-Flash-Next (Gyro) Gyro-S 58 GB 28 GB fitsQwen3.8-Flash-Next (Gyro) Gyro-M 92 GB 35 GB fits- Qwen3.8-Flash-Next (NVFP4) NVFP4 128 GB 69 GB does not fit (can still try) · use- Qwen3.8-Flash-Next (NVFP4) NVFP4-NVIDIA 128 GB 69 GB does not fit (can still try) · use- Qwen3.8-Flash-Next (NVFP4) NVFP4-QAD 102 GB 69 GB does not fit (can still try) · use+ Qwen3.8-Flash-Next (NVFP4) NVFP4 128 GB 69 GB too big (can still try) · SSD 7 GB+ Qwen3.8-Flash-Next (NVFP4) NVFP4-NVIDIA 128 GB 69 GB too big (can still try) · SSD 7 GB+ Qwen3.8-Flash-Next (NVFP4) NVFP4-QAD 102 GB 69 GB too big (can still try) · SSD 7 GBdiff --git a/docs/terminal-previews/launch-step2-folder.txt b/docs/terminal-previews/launch-step2-folder.txtindex 91d366a26..16544d7fc 100644--- a/docs/terminal-previews/launch-step2-folder.txt+++ b/docs/terminal-previews/launch-step2-folder.txt@@ -5,7 +5,7 @@────────────────────────────────────────────────────────────────────────────────────────────────────────────Where should the model files be stored?Needs about 84 GB (the download plus working room).- Free space in that folder's drive: 52 GB Not enough for 84 GB. You can still go on, but the download is+ Free space in that folder's drive: 47 GB Not enough for 84 GB. You can still go on, but the download islikely to stop; pick a bigger drive.┌──────────────────────────────────────────────────────────────────────────────────────────────────────────┐│ C:\Users\AI-Server\Documents\Strata │diff --git a/docs/terminal-previews/launch-twocards-start.txt b/docs/terminal-previews/launch-twocards-start.txtindex 2f2f35293..54af704b2 100644--- a/docs/terminal-previews/launch-twocards-start.txt+++ b/docs/terminal-previews/launch-twocards-start.txt@@ -1,7 +1,7 @@# strata launch preview: start screen with saved presets# (text of the screen as drawn at the size named in tests/terminal/launch_previews.py)- STRATA Launch NVIDIA RTX A4000 16 GB x2 · 64 GB RAM · 52 GB free+ STRATA Launch NVIDIA RTX A4000 16 GB x2 · 64 GB RAM · 47 GB free────────────────────────────────────────────────────────────────────────────────────────────────────────────Start from a saved preset@@ -31,6 +31,6 @@────────────────────────────────────────────────────────────────────────────────────────────────────────────- This computer: NVIDIA RTX A4000 16 GB x2 · 64 GB RAM · 52 GB free. Preset: model, context, parallel+ This computer: NVIDIA RTX A4000 16 GB x2 · 64 GB RAM · 47 GB free. Preset: model, context, parallelrequests, tuned settings, access. Shared with the Strata app.↑↓ choose Enter launch e edit preset d delete ? help q quitdiff --git a/strata_cli/launch/catalog.py b/strata_cli/launch/catalog.pyindex b0420919c..ebfe0fcb8 100644--- a/strata_cli/launch/catalog.py+++ b/strata_cli/launch/catalog.py@@ -70,7 +70,7 @@ def verdict(on_this_pc: str | None) -> tuple[str, str]:if not text:return TIGHT, "not checked (this computer's memory could not be read)"if low.startswith("does not fit"):- return NO, "does not fit (can still try)"+ return NO, "too big (can still try)"if "low-ram mode" in low:return FITS, "fits, a bit slower"if "part of its experts" in low or "read from the ssd" in low:diff --git a/strata_cli/launch/cmd.py b/strata_cli/launch/cmd.pyindex 5dc24d8a4..0412429b2 100644--- a/strata_cli/launch/cmd.py+++ b/strata_cli/launch/cmd.py@@ -234,7 +234,7 @@ def apply_defaults(s: S.LaunchSettings, catalog) -> tuple[S.LaunchSettings, listif rows:r = rows[0]if r.verdict == "no":- notices.append(S.Notice("warn", f"{r.title} does not fit this computer ({r.words}).", "It can still start; expect it to be slow or to run out of memory."))+ notices.append(S.Notice("warn", f"{r.title} does not fit this computer.", "It can still start; expect it to be slow or to run out of memory."))elif r.verdict == "tight":notices.append(S.Notice("warn", f"{r.title} is a tight fit: {r.words}.", "Close other programs before you start."))if s.vision != "off" and not r.images:diff --git a/strata_cli/launch/screens.py b/strata_cli/launch/screens.pyindex 510ff56db..f090b684e 100644--- a/strata_cli/launch/screens.py+++ b/strata_cli/launch/screens.py@@ -353,8 +353,8 @@ class ConfirmScreen(ModalScreen):class ModelScreen(Base):step_id = "model"hints = "↑↓ choose Enter next o use my own files Esc back ? help"- help_text = ("Pick the model. 'Needs' is the RAM + graphics memory the model wants (the app's figure); 'uses SSD N GB' is the part that does not fit and is read from the SSD while it answers; the n-gram table stays on "- "your SSD and is not counted. 'Does not fit' is a warning, not a block: you can still try it.")+ help_text = ("Pick the model. 'Needs' is the RAM + graphics memory the model wants (the app's figure); 'SSD N GB' is the part that does not fit and is read from the SSD while it answers; the n-gram table stays on "+ "your SSD and is not counted. 'Too big' is a warning, not a block: you can still try it.")BINDINGS = Base.BINDINGS + [Binding("o", "own", "Own files", show=False)]@@ -390,7 +390,7 @@ class ModelScreen(Base):col = {"fits": "", "tight": WARN, "no": BAD}[r.verdict]tag = (" (recommended)" if r.recommended else "") + (" (installed)" if r.installed else "") + (" [experimental]" if r.experimental else "")need = f"{r.memory_gb:.0f} GB" if r.memory_gb else "-"- ssd = f" · uses SSD {r.spill_gb:.0f} GB" if r.spill_gb >= 1 else ""+ ssd = f" · SSD {r.spill_gb:.0f} GB" if r.spill_gb >= 1 else ""items.append(Item([r.title, f"{r.download_gb:.0f} GB", need, Cell((r.words.split(",")[0].split(":")[0] if ssd and "SSD" in r.words else r.words) + ssd + tag, col)], key=(r.family, r.model)))if (s.family, s.model) == (r.family, r.model):sel = idiff --git a/tests/terminal/test_launch_tui.py b/tests/terminal/test_launch_tui.pyindex 330350355..c6c41dd18 100644--- a/tests/terminal/test_launch_tui.py+++ b/tests/terminal/test_launch_tui.py@@ -230,9 +230,9 @@ class ModelStep(Base):self.assertIn("44 GB", line) # the app's figure: experts 42.9 + 1 GB, RAM + VRAM togetherself.assertIn("(recommended)", line)self.assertIn("› ", line) # the cursor starts on the recommendation- self.assertIn("does not fit (can still try)", t) # NVFP4: a warning, not a block+ self.assertIn("too big (can still try)", t) # NVFP4: a warning, not a blockbig = next(l for l in t.splitlines() if "UD-Q5_K_XL" in l)- self.assertIn("uses SSD", big) # a model bigger than RAM says how much is read from the SSD, as the app does+ self.assertIn("SSD ", big) # a model bigger than RAM says how much is read from the SSD, as the app doesself.assertIn("The n-gram table stays on your SSD", t)self.run_app(case)