Paste box goes live: fine-tuned spec parser fills the form from messy text

.gitattributes

@@ -34,3 +34,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 model/speed_model.skops filter=lfs diff=lfs merge=lfs -text
+model/spec_lora/tokenizer.json filter=lfs diff=lfs merge=lfs -text

app.py

@@ -75,6 +75,14 @@ def api_lookup(payload: LookupIn):
     return lookup(p.get("repo", ""), p, spec_from_payload(p))
 
 
+@app.api(name="parse")
+def api_parse(text: str = "") -> dict:
+    """Messy machine description -> form fields, via the fine-tuned spec
+    parser (cn0303/fitcheck-spec-parser). ZeroGPU via the Gradio queue."""
+    from spec_brick import parse_specs
+    return parse_specs(text)
+
+
 @app.api(name="ask")
 def api_ask(question: str, facts: str = "") -> dict:
     """Plain-English follow-up, grounded in the facts /api/advise returned.

model/metrics.json

@@ -7,5 +7,15 @@
   "baseline_mape_pct": 28.1,
   "baseline_mae_tps": 11.63,
   "model_mape_on_bw_known_pct": 17.5,
-  "model_mae_on_bw_known_tps": 9.55
+  "model_mae_on_bw_known_tps": 9.55,
+  "envelope": {
+    "bytes_gb": [
+      0.84,
+      9.85
+    ],
+    "eff_bw": [
+      68.0,
+      3350.0
+    ]
+  }
 }

model/spec_lora/README.md

@@ -0,0 +1,62 @@
+---
+license: apache-2.0
+base_model: unsloth/Qwen3-1.7B
+library_name: peft
+pipeline_tag: text-generation
+tags: [lora, sft, structured-extraction, hardware-specs, qwen3, unsloth]
+---
+
+# FitCheck spec parser (Qwen3-1.7B LoRA)
+
+Turns messy human descriptions of computers — "my dad's old Dell, i5, 16 gigs,
+some nvidia card" — into the structured spec JSON used by
+[FitCheck](https://huggingface.co/spaces/build-small-hackathon/FitCheck), the
+honest "what AI can your computer run" advisor. This powers its paste box.
+
+The one rule that matters: **missing information becomes `null`, never a
+guess.** The model is trained and gated specifically against inventing specs.
+
+## Training data: grounded, not synthetic-echo
+
+Labels are never model-generated: every training example starts from a **real
+machine** (GPUs + VRAM from vendor spec tables, 212 cards + Apple chips); only
+the phrasing varies, across ~24 registers mimicking how people actually write
+(casual chat, dxdiag dumps, Task Manager paste, seller listings, consoles,
+comparisons, half-remembered specs, several languages). ~39% of examples have
+no GPU to extract — the don't-invent cases. Trained with Unsloth (bf16 LoRA,
+completion-only loss) on a single RTX 5090 laptop.
+
+## Evaluation: human-written text only
+
+Evaluated on a 45-example **human-written dev set** (never generator output;
+multilingual, consoles, buying-intent traps, pure refusals). The builder
+iterated against this set, so these are **dev numbers** — optimistically
+biased by adaptive iteration, and labelled as such:
+
+| round | field accuracy | invented-field rate (hallucination) |
+|---|---|---|
+| 1 | 77.3% | 32.5% |
+| 3 (answer-only loss + explicit rules) | 85.8% | 12.0% |
+| 5 (final) | **91.6%** | **1.2%** |
+
+A **sealed test set** (written by people who never saw the training data,
+evaluated exactly once, builder-blind) is pending; its result will be added
+here unedited when run. Ship gate: beat the base model zero-shot AND keep the
+invented-field rate under 5% — passed on dev.
+
+## Output schema
+
+```json
+{"computer": "Windows laptop|Windows desktop|Mac|Linux PC|Mini PC / Raspberry Pi|null",
+ "ram_gb": "number|null", "provider": "nvidia|amd|apple|intel|none|null",
+ "gpu": "string|null", "vram_gb": "number|null"}
+```
+
+Notable learned rules: `"none"` only when the text says there's no graphics
+card (unknown → null); a series alone ("gtx") is a provider, not a GPU; a
+stated VRAM figure beats the model's knowledge of that card; dxdiag's
+"Display Memory" is not system RAM; "8gb dev kit" on a Jetson is unified RAM,
+not VRAM; two machines compared → extract nothing.
+
+Part of the FitCheck project (Build Small hackathon): a deterministic engine
+does the math; small models appear only where they earn their place.

model/spec_lora/adapter_config.json

@@ -0,0 +1,52 @@
+{
+  "alora_invocation_tokens": null,
+  "alpha_pattern": {},
+  "arrow_config": null,
+  "auto_mapping": {
+    "base_model_class": "Qwen3ForCausalLM",
+    "parent_library": "transformers.models.qwen3.modeling_qwen3",
+    "unsloth_fixed": true
+  },
+  "base_model_name_or_path": "unsloth/Qwen3-1.7B",
+  "bias": "none",
+  "corda_config": null,
+  "ensure_weight_tying": false,
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0,
+  "lora_ga_config": null,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "peft_version": "0.19.1",
+  "qalora_group_size": 16,
+  "r": 16,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "up_proj",
+    "down_proj",
+    "q_proj",
+    "v_proj",
+    "o_proj",
+    "k_proj",
+    "gate_proj"
+  ],
+  "target_parameters": null,
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_bdlora": null,
+  "use_dora": false,
+  "use_qalora": false,
+  "use_rslora": false
+}

model/spec_lora/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7da9d72df630128f602da0724e78f33b98406752ba821b998ea2be521f90e6f4
+size 69782384

model/spec_lora/chat_template.jinja

@@ -0,0 +1,98 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0].role == 'system' %}
+        {{- messages[0].content + '\n\n' }}
+    {%- endif %}
+    {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0].role == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for forward_message in messages %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- set message = messages[index] %}
+    {%- set current_content = message.content if message.content is not none else '' %}
+    {%- set tool_start = '<tool_response>' %}
+    {%- set tool_start_length = tool_start|length %}
+    {%- set start_of_message = current_content[:tool_start_length] %}
+    {%- set tool_end = '</tool_response>' %}
+    {%- set tool_end_length = tool_end|length %}
+    {%- set start_pos = (current_content|length) - tool_end_length %}
+    {%- if start_pos < 0 %}
+        {%- set start_pos = 0 %}
+    {%- endif %}
+    {%- set end_of_message = current_content[start_pos:] %}
+    {%- if ns.multi_step_tool and message.role == "user" and not(start_of_message == tool_start and end_of_message == tool_end) %}
+        {%- set ns.multi_step_tool = false %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {%- set content = message.content %}
+        {%- set reasoning_content = '' %}
+        {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+            {%- set reasoning_content = message.reasoning_content %}
+        {%- else %}
+            {%- if '</think>' in message.content %}
+                {%- set content = (message.content.split('</think>')|last).lstrip('\n') %}
+                {%- set reasoning_content = (message.content.split('</think>')|first).rstrip('\n') %}
+                {%- set reasoning_content = (reasoning_content.split('<think>')|last).lstrip('\n') %}
+            {%- endif %}
+        {%- endif %}
+        {%- if loop.index0 > ns.last_query_index %}
+            {%- if loop.last or (not loop.last and reasoning_content) %}
+                {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message.role + '\n' + content }}
+            {%- endif %}
+        {%- else %}
+            {{- '<|im_start|>' + message.role + '\n' + content }}
+        {%- endif %}
+        {%- if message.tool_calls %}
+            {%- for tool_call in message.tool_calls %}
+                {%- if (loop.first and content) or (not loop.first) %}
+                    {{- '\n' }}
+                {%- endif %}
+                {%- if tool_call.function %}
+                    {%- set tool_call = tool_call.function %}
+                {%- endif %}
+                {{- '<tool_call>\n{"name": "' }}
+                {{- tool_call.name }}
+                {{- '", "arguments": ' }}
+                {%- if tool_call.arguments is string %}
+                    {{- tool_call.arguments }}
+                {%- else %}
+                    {{- tool_call.arguments | tojson }}
+                {%- endif %}
+                {{- '}\n</tool_call>' }}
+            {%- endfor %}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+    {%- if enable_thinking is defined and enable_thinking is false %}
+        {{- '<think>\n\n</think>\n\n' }}
+    {%- endif %}
+{%- endif %}

model/spec_lora/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7430e9138b76e93fb6f93462394d236b411111aef53cb421ba97d2691040cca
+size 11423114

model/spec_lora/tokenizer_config.json

@@ -0,0 +1,233 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "is_local": false,
+  "model_max_length": 40960,
+  "pad_token": "<|PAD_TOKEN|>",
+  "padding_side": "left",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151665": {
+      "content": "<tool_response>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151666": {
+      "content": "</tool_response>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151667": {
+      "content": "<think>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151668": {
+      "content": "</think>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": false
+    },
+    "151669": {
+      "content": "<|PAD_TOKEN|>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    }
+  }
+}

model/speed_model.skops

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:70249e8241e24de5202af7ba82e911b9dbfaed06c19be920d6b6fd2873bef58b
+oid sha256:43f0f404f6cd13cfad072c764bcfebb9775cd56b994bbd65669185ce04228319
 size 1658817

requirements.txt

@@ -14,3 +14,4 @@ accelerate                # device placement / efficient loading
 einops                    # required by the kernels-community mamba-ssm kernel
 skops                     # safe loading of the trained speed predictor
 xgboost                   # the speed predictor's runtime (engine/speed.py)
+peft                      # spec-parser LoRA adapter loading

spec_brick.py

@@ -0,0 +1,96 @@
+"""
+The spec-parser brick: messy human text -> the advisor's form fields.
+
+Serves cn0303/fitcheck-spec-parser (Qwen3-1.7B + LoRA, trained in this repo —
+see scripts/train_spec_lora.py and the model card for the honest eval). Same
+serving pattern as the narrator: lazy load inside @spaces.GPU, loud errors,
+no fake fallbacks. Missing info comes back null — the model is specifically
+gated against inventing specs.
+"""
+
+import json
+import re
+import sys
+
+from model_brick import _should_load
+
+ADAPTER_ID = "cn0303/fitcheck-spec-parser"
+BASE_ID = "unsloth/Qwen3-1.7B"
+
+# MUST stay in sync with scripts/build_spec_dataset.py (the training prompt).
+SYSTEM_PROMPT = """\
+You turn a person's description of their computer into JSON for a hardware checker.
+Output ONLY a JSON object with exactly these fields:
+{"computer": "Windows laptop"|"Windows desktop"|"Mac"|"Linux PC"|"Mini PC / Raspberry Pi"|null,
+ "ram_gb": number|null, "provider": "nvidia"|"amd"|"apple"|"intel"|"none"|null,
+ "gpu": string|null, "vram_gb": number|null}
+Rules:
+- Extract ONLY what the text states or directly implies. Anything not stated is null. Never guess or invent a spec.
+- "provider": "none" ONLY when the text says there is no separate graphics card (e.g. "no GPU", "integrated only"). Graphics simply not mentioned or unknown -> null.
+- "gpu" must be a specific model (e.g. "RTX 3060"). A brand or series alone ("geforce", "gtx", "radeon") is NOT a gpu -> set provider, leave gpu null.
+- If the text describes two or more different machines or a choice between them, every field is null."""
+
+FIELDS = ("computer", "ram_gb", "provider", "gpu", "vram_gb")
+
+_GENERATE = None
+_state = {"tok": None, "model": None}
+
+if _should_load():
+    try:
+        import spaces
+        import torch
+        from transformers import AutoModelForCausalLM, AutoTokenizer
+
+        def _load():
+            from peft import PeftModel
+            tok = AutoTokenizer.from_pretrained(ADAPTER_ID)
+            model = AutoModelForCausalLM.from_pretrained(BASE_ID, dtype=torch.bfloat16)
+            model = PeftModel.from_pretrained(model, ADAPTER_ID)
+            _state["tok"] = tok
+            _state["model"] = model.to("cuda").eval()
+
+        @spaces.GPU(duration=60)
+        def _generate(text: str) -> str:
+            if _state["model"] is None:
+                _load()
+            tok, model = _state["tok"], _state["model"]
+            msgs = [{"role": "system", "content": SYSTEM_PROMPT},
+                    {"role": "user", "content": text}]
+            kw = dict(add_generation_prompt=True, return_tensors="pt", return_dict=True)
+            try:
+                inputs = tok.apply_chat_template(msgs, enable_thinking=False, **kw)
+            except TypeError:
+                inputs = tok.apply_chat_template(msgs, **kw)
+            inputs = inputs.to("cuda")
+            n = inputs["input_ids"].shape[1]
+            with torch.no_grad():
+                out = model.generate(**inputs, max_new_tokens=96, do_sample=False,
+                                     pad_token_id=tok.eos_token_id)
+            return tok.decode(out[0][n:], skip_special_tokens=True).strip()
+
+        _GENERATE = _generate
+    except Exception as e:  # noqa: BLE001
+        print(f"[FitCheck] spec parser unavailable: {e!r}", file=sys.stderr, flush=True)
+
+
+def parse_specs(text: str) -> dict:
+    """Returns the parsed fields, or {error} — never invented content."""
+    text = (text or "").strip()
+    if not text:
+        return {"error": "Nothing to parse — paste or type a description first."}
+    if len(text) > 4000:
+        text = text[:4000]
+    if _GENERATE is None:
+        return {"error": "The spec parser model isn't loaded in this environment."}
+    try:
+        raw = _GENERATE(text)
+    except Exception as e:  # noqa: BLE001
+        return {"error": f"Spec parser failed: {e}"}
+    m = re.search(r"\{.*\}", raw, re.DOTALL)
+    if not m:
+        return {"error": f"The parser didn't return JSON. Raw output: {raw[:200]}"}
+    try:
+        obj = json.loads(m.group(0))
+    except json.JSONDecodeError:
+        return {"error": f"The parser returned malformed JSON: {m.group(0)[:200]}"}
+    return {f: obj.get(f) for f in FIELDS}

static/app.js

@@ -567,6 +567,61 @@ async function drawRoofline(speed) {
   host.innerHTML = s;
 }
 
+// ---- Paste box: the fine-tuned spec parser fills the form -----------------
+async function parsePaste() {
+  const text = $("#paste").value.trim();
+  const hint = $("#parse-hint");
+  if (!text) { hint.textContent = "Paste or type something first."; return; }
+  hint.textContent = "Reading it… (first use after a quiet spell wakes the model, up to a minute)";
+  try {
+    const client = await getClient();
+    const r = await client.predict("/parse", { text });
+    const d = Array.isArray(r.data) ? r.data[0] : r.data;
+    if (d.error) { hint.textContent = d.error; return; }
+    applyParsed(d, hint);
+  } catch (e) {
+    hint.textContent = `Parser unavailable: ${e && e.message ? e.message : e}`;
+  }
+}
+
+function applyParsed(d, hint) {
+  const got = [];
+  if (d.computer) {
+    state.computer = d.computer;
+    setActive("#computer-seg", d.computer);
+    syncProviderForComputer();
+    got.push(d.computer);
+  }
+  if (d.provider) {
+    state.provider = d.provider;
+    setActive("#provider-seg", d.provider);
+    fillGpu();
+    got.push(d.provider.toUpperCase());
+  }
+  if (d.gpu) {
+    const sel = $("#gpu");
+    const want = String(d.gpu).toLowerCase().replace(/\b(nvidia|geforce|amd|radeon|intel)\b/g, "").trim();
+    const match = [...sel.options].find(o => o.value.toLowerCase().includes(want));
+    if (match) { sel.value = match.value; got.push(match.value); }
+    else if (d.vram_gb) { $("#vram").value = d.vram_gb; got.push(`${d.gpu} (${d.vram_gb} GB)`); }
+    else got.push(d.gpu);
+  } else if (d.vram_gb) {
+    $("#vram").value = d.vram_gb;
+    got.push(`${d.vram_gb} GB VRAM`);
+  }
+  if (d.ram_gb) {
+    const ram = $("#ram");
+    const opt = [...ram.options].map(o => parseFloat(o.value))
+      .reduce((a, b) => Math.abs(b - d.ram_gb) < Math.abs(a - d.ram_gb) ? b : a);
+    ram.value = `${opt} GB`;
+    got.push(`${d.ram_gb} GB RAM`);
+  }
+  hint.textContent = got.length
+    ? `Understood: ${got.join(", ")}. Anything you didn't mention stayed blank — confirm and press Check.`
+    : "Couldn't find any specs in that text — nothing was filled in (the parser doesn't guess).";
+  maybeLiveUpdate();
+}
+
 // ---- Follow-up: the model brick (grounded explainer) ---------------------
 function wireAsk() {
   const input = $("#ask-input"), send = $("#ask-send");
@@ -656,6 +711,7 @@ function init() {
   ["#ram","#gpu","#vram","#custom-uc","#repo-check"].forEach(s => { const el = $(s); if (el) el.addEventListener("change", maybeLiveUpdate); });
   $("#paste").addEventListener("input", maybeLiveUpdate);
   $("#check-btn").addEventListener("click", check);
+  const pb = $("#parse-btn"); if (pb) pb.addEventListener("click", parsePaste);
   fillGpu();
   $("#find-specs-body").innerHTML = findSpecsText();
   detectHardware();

static/index.html

@@ -90,6 +90,8 @@
               <div class="field" style="margin-bottom:0">
                 <span class="label">Or paste / describe your specs</span>
                 <textarea id="paste" placeholder="Paste output from 'dxdiag' or 'Task Manager → Performance', or just describe it: 'Dell XPS, RTX 3050, 16GB'…"></textarea>
+                <button class="parse-btn" id="parse-btn">Fill the form from this</button>
+                <div class="hint" id="parse-hint">A small model fine-tuned for this reads your text and fills the form. It never guesses: anything you didn't say stays blank.</div>
               </div>
             </div>
           </details>

static/style.css

@@ -240,6 +240,15 @@ details.disc > summary:hover { color: var(--text-primary); }
 .disc-body dt { font-weight: 700; color: var(--text-primary); font-family: var(--font-head); }
 .disc-body dd { color: var(--text-muted); }
 
+.parse-btn {
+  margin-top: var(--s-2); width: 100%;
+  background: var(--accent-soft); border: 1px solid var(--accent);
+  color: var(--text-primary); border-radius: var(--r-sm);
+  padding: 9px 12px; font-size: 13.5px; font-weight: 600;
+  transition: filter .15s, transform .15s;
+}
+.parse-btn:hover { filter: brightness(1.15); transform: translateY(-1px); }
+
 /* Primary CTA */
 .cta {
   width: 100%; border: none; border-radius: var(--r-md);